Screen Grid Experimentation for AM

Discussion in 'Amplitude Modulation' started by AC0OB, Apr 28, 2021.

ad: L-HROutlet
ad: l-rl
ad: L-Geochron
ad: Left-3
ad: abrind-2
ad: L-MFJ
ad: Left-2
  1. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    Further experimentation with my HT-40 Turbo has yielded some (to me) interesting results. My HT-40 turbo has a 6DE7 cathode follower feeding the screen of the 6DQ5 final, with Vp of ~ 700 volts. I went back to experiment first with a CC time constant circuit, but with a resistance lower than found in the old novice transmitter designs, Page 1.

    Most designs had the cathode referenced to ground with the CC time constant circuit in series between the cathode and the final's screen. The oem CC circuits used a resistor > 10k with the capacitance selected to yield a response down to about 30 Hz. Since capacitance was expensive back then, smaller values of capacitance meant that the higher value of the resistance dropped a lot of voltage across it, resulting in low power with no modulation. The quiescent Vsg was around 65V, so when the peak-to-peak audio was > 170V, the CC circuit attempted to drive the signal below ground, resulting in a flattening of the composite signal at the cathode. This guaranteed that the modulation percentage was only about 80-85%.

    So I opted for a 3.3k resistor with a larger capacitance of 2.2 uF, and a pulldown to the bias voltage already available. This allowed > 99% modulation without the flattening of the composite signal and a higher carrier power without modulation. IMHO, this is a decent design for most Novice SG transmitters with CC.

    Not content with that, I decided to investigate and analyze the Continental 316B series of SG modulated Broadcast transmitters to determine how they developed their so-called, "Constant Carrier" system without massive resistors and expensive capacitors, Page 3, simplified schematic.

    What they did was to use a feedback system in which they sampled the audio driver's cathode voltage and then corrected the cathode-follower's grid voltage with an 807 error voltage amplifier. The 4CX5000's screen grid required between 350 and 550 volts for screen modulation at a Vp of 7.5kV.

    Before I could duplicate this somehow, I needed to know how the 6DE7 or 6EW7 Unit 2 reacted to various grid voltages verses positive cathode voltages with a constant plate voltage, so I characterized it, Page 6. This shows that this series of dissimilar triode tubes is pretty darn linear wrt DC.

    Due to the fact that only a range of negative grid voltages could give us the required range of suitable positive cathode voltages, the circuit of Page 4 was built. Now we have modulation up to 125%, greater efficiency (56% with modulation) and no need for the RC/CC system. Power output can be varied between 15 and 50 Watts without having to change the pi-net loading capacitance.

    The disadvantages are that we still need a bias voltage (and the extra space for the transformer and circuit components), a P-MOSFET for error control (getting harder to find), and increased circuit complexity over the simple and economical OEM circuit.


    Attached Files:

    Last edited: Apr 28, 2021

Share This Page