Negative half cycle modulation transformer loading?

Discussion in 'Amplitude Modulation' started by W1BR, Sep 22, 2019.

ad: L-HROutlet
ad: l-rl
ad: Subscribe
ad: MessiPaoloni-1
ad: L-MFJ
ad: K5AB-Elect-1
ad: Left-2
  1. W1BR

    W1BR Ham Member QRZ Page

    Was this a common practice? I saw it suggested as a modification for a Viking transmitter--circuit changes had steering diodes that would place loading resistors across the modulated B plus on negative going modulation excursions. Was this really needed in the Johnson transmitters? Good idea, or waste of time?
     
  2. AC0OB

    AC0OB Subscriber QRZ Page

    When I do upgrades to Valiants, Viking II's, and Apache's I use a circuit idea from Bonomo that made sense (at least to me:)).

    Basically, it loads the modulation transformer secondary when the negative excursion reaches a certain level in order to avoid a destructive V = - L*di/dt spike.


    Pheel
     

    Attached Files:

    Last edited: Sep 22, 2019
  3. AC0OB

    AC0OB Subscriber QRZ Page


    Addendum:

    I also add this circuit to DX-100's as well.


    Pheel
     
  4. K4KYV

    K4KYV Subscriber QRZ Page

    I would avoid it.

    Loading down the modulator on the negative half of each cycle causes the impedance of the load to vary over the audio cycle, inherently distorting the waveform. Plus, the rectified audio voltage appears across the resistor, wasting audio power by converting it to heat per Ohm's law. Audio power is too difficult and expensive to generate and then burn up in a resistor, and the waveform distortion caused by this circuit produces spurious sideband products, AKA splatter or "wide" signals.

    As I recall, the Viking II modulator is just barely capable of modulating the carrier 100% in either direction, and certainly has little or no audio power left to spare to heat up a resistor.

    This circuit is known as "negative cycle loading". A similar, more sophisticated version known as "ultramodulation" uses three diodes and a tap on one of the resistors to act as a voltage divider on the negative half of the cycle.

    The purported function of these circuits is to expand the positive peaks and limit negative peaks, allegedly allowing more sideband power with posititive peaks exceeding 100% while avoiding breaking base line with negative cycle overmodulation. First of all, these schemes rely on having a modulator with peak output power capability in excess of what is needed to modulate 100%. Most popular plate modulated amateur transmitters do not fit the bill. Nor do military surplus transmitters like the BC-610 and T-368. Usually, the modulation transformer has enough step-down ratio to limit modulation capability just barely to 100%. This is intended as a protection feature, to make it impossible to modulate very much in excess of 100% if even that much.

    Negative cycle loading and ultramodulation are totally useless in a transmitter that is incapable of modulating more than 100% to begin with.

    Here's another circuit using a resistor and rectifier, that has some merit if the modulator has some reserve headroom allowing it to modulate substantially beyond 100%. See attached schematic. It maintains the proper load on the modulation transformer in the event of negative peak overmodulation. Using a MV rectifier tube like the 866-A as the diode, allows it to act as a visual overmodulation indicator as well. The resistor should have the same resistance in ohms as the modulation impedance of the RF final: plate voltage (volts) รท plate current (amperes). The wattage rating of the resistor is small, 10% of the DC input to the final is probably enough, since the portion of the audio cycle in excess of 100% negative overmodulation with a normal modulator is very small.

    For further explanation, click on this link and scroll to message #21.
     

    Attached Files:

    K5UJ likes this.
  5. AC0OB

    AC0OB Subscriber QRZ Page

    I would not be so quick to dismiss Bonomo's soft knee limiter.

    I reproduced K4KYV's (equivalent) circuit on page 2 of the pdf file and I don't see any difference in operation. The only difference is in the implementation.

    I dug out my back issues of ER magazine and issues # 110 and #111 and the author is Thomas Bonomo, K6AD, and his articles are entitled, "Viking II Modifications - Revisiting a Legend."

    Issue # 111 page 32 is where he has the schematic of the soft knee negative limiter.

    On Pages 34-35 he describes the complete operation of the Limiter and makes no claims of reduced distortion for any limiter, whether it be positive or negative limiters.

    He says on Page 35, "...A clipped waveform contains the same harmonics, regardless of whether the positive or negative peaks are clipped..."


    "...Only to the extent that the peak limiter produces 'softer' clipping of the negative modulation peaks will splatter be reduced. While the 807 modulator in the Viking II doesn't have enough extra power to drive positive peaks beyond much beyond 100%, the negative peak limiter is still quite useful in protecting the modulation transformer and should be installed for this reason alone..."

    So he makes no claims other than soft negative peak limiting and protection of the modulation transformer.

    He never states anywhere that this circuit will increase positive modulation percentage, nor does he claim Viking era transmitters can produce much more than 105% positive modulation.


    Pheel

     

    Attached Files:

    Last edited: Sep 22, 2019
  6. AC0OB

    AC0OB Subscriber QRZ Page


    Correction: I had the polarity of D1=866 flipped on page 2.
     

    Attached Files:

  7. K5UJ

    K5UJ Ham Member QRZ Page

    Unless I'm mistaken what's being debated is the two ways of performing neg. cycle loading. I agree with Don: The bad way is to parallel reversed diodes and a resistor with the mod. transformer secondary. Then the primary drives the final B+ below its dead carrier voltage, the resistor appears across the secondary winding, which changes its load Z plus the resistor is dissipating a lot of heat, all points Don made. The good way in my opinion, is to run the reverse diode string, using enough so their total PIV is comfortably above the peak positive modulated B+, from a point on the B+ line between the mod. trans. and the cold side of the plate choke, to a resistor that has a value equal to the final PA anode impedance, with the other side of the resistor grounded. Now the diodes conduct only when the B+ is driven to zero, then there's a resistor of several thousand ohms to act as a load so the mod. transformer has something other than infinite impedance. That's the theory anyway. I've tried it and at the least, it doesn't seem to hurt anything, and I've driven my 100 w. rig to cut off and the mod. iron is still okay (so far). I'm using regular power supply silicon diodes and I am not sure they are best for this application.

    Something I've never tried but wondered about, is putting a positive v. of say 70 v. at the junction between the last diode and the resistor, which I think would cause them to conduct once the B+ is modulated below 70 v. in this example.
     
  8. AC0OB

    AC0OB Subscriber QRZ Page

    There are many ways to implement a negative voltage limiter and except for the implementation, the action of the Bonomo limiter and Don's limiter is the same, assuming I understand Don's limiter. Don's limiter uses an external bias voltage to set the onset of limiting as to when the 866 diode tube conducts, whereas here the limiting voltage is set by a set of zeners.

    Sheet three is a simplified diagram of the limiter.

    D1 is a 180V zener, D2 and D3 are 200 volt zeners. D1 - D3 form a series zener string set to conduct at 580V.

    D4 is a switching diode oriented to conduct only when the voltage at the top of the modulation transformer starts to exceed minus 580.7 volts. D4 could as easily be placed between the top of the transformer and the loading resistor or anywhere in between; it doesn't matter.

    So when the voltage at the top of the modulation transformer starts to exceed minus 580.7 volts (with respect to the bottom rail of the transformer's secondary), current starts flowing through the 1.2k load resistor which presents a constant load ONLY during the negative excursion of the secondary waveform.

    This effectively protects the modulation secondary from any negative transient kickback voltages (V = - L*di/dt) greater than 580.7 volts across the modulation secondary. Any energy resulting from high negative voltage transients is dissipated in the resistor.

    I.e., a finite load is presented to the secondary when voltages > - 580.7 Volts go negative.


    Pheel
     

    Attached Files:

    Last edited: Sep 23, 2019
  9. N2DTS

    N2DTS Ham Member QRZ Page

    I tried this many years ago, with an adjustable power supply to set the bias (limiting point).
    I used the correct value of resistor and three 14 kv 1 amp diode bricks.
    My modulators can produce as much audio as power output of the RF deck, or more.
    I got reports my signal was very wide and scrapped the system.
    I was told by someone long ago it was a problem with the diodes not reacting fast enough to the high frequency stuff in the audio.

    Another thought, unless you are running things on the edge (maybe some stock rigs do?), over modulating seems unlikely to blow up mod transformers.
    For low power rigs (100 watts or less) I always tended to put a lot of audio on the carrier and never blew a mod transformer.
    The only time I ever blew a mod transformer (a few in a row) was using old rusty open frame 1500 volt RCA mod iron at 2500 volts.
    You can over modulate them fine at 1500 volts...

    High power broadcast rigs may be different?
     
  10. AC0OB

    AC0OB Subscriber QRZ Page

    I didn't know we were debating or competing. W1BR asked about negative half-cycle loading. I presented a soft knee limiter circuit that had been developed years ago and one that has worked for me in countless upgrades, and saved many an un-obtanium modulation transformer from dying.

    There may be other methods that do half-cycle loading to avoid infinite impedance, high voltage transients on negative excursions, but Bonomo's circuit seems to be the simplest and most economical way to implement it.


    Pheel
     

Share This Page