now lets talk abt something really important - "self modulation" of screens

[ke7trp]

Sounds good..  Post some pictures.  Mine is now switchable.  In the SSB mode, I have a 0 to 1000Volt supply on a variac with protection and metering. When in AM, It switches in a Dropping resistor so the Screen is modulated. When its finished, I will post pictures and specs.

C

[Ed/KB1HYS]

I had thought about switching the choke out, but I don't need a half-gallon CW rig, and trying to use the 4-250 as a linear Amp at the power & voltages I have just ain't going to get the job done. So Keep It Simple and just build an 'phone rig. 160, 80 and 40 meters maybe If I can swing the tank circuits right.

[Opcom]

I do just about all my testing and transmitting into a dummy load, so I have no one to give any reports but I have done a lot of experimenting.

The rig I have been playing with uses a 4-1000 PA and 3-500Z modulators running 2800VDC. It has a screen supply with 10% regulation and an output impedance of a 2uF capacitor, so it is not the greatest. It has choke modulation through a  8H choke. With this setup there was no problem with 100% modulation and voice sounded fine but at low frequencies according to test equipment the linearity was not very good. It was not the modulator's fault, nor really the low-ish standing voltage on the 4-1000. The choke did not do as well as maybe it should have. The screen current was about 140mA and the voltage was 500 so the screen impedance was about 3500 ohms. The  frequency where the choke impedance equal the screen impedance was about 59Hz and as the frequency of the modulating signal got below 200Hz or so the screen modulation became less effective. A higher value would have been better, like 20H, so the Z of the choke would have been much more than the Z of the screen grid at the 59Hz point.

In some experiments the best results from this transmitter were gotten (in my case) by using a screen modulation winding. The transmitter did not have one, so I used a smaller modulation transformer and capacitively coupled it to the modulated plate voltage through a series resistor so that the correct voltage swing would be present on the primary of the smaller transformer to reflect the degree of screen modulating voltage needed for the screen grid. That made it more complicated but I think eventually I could mount this mod xfmr in place of the 8H choke because it is a 300mA unit and 2x bigger than necessary.

The additional transformer was a military one designed for a pair of 4X150's modulating three 4X150's at 1700V and is rated 125W with a 2db response from 200Hz to 20,000Hz and in the testing the response at the low end was good to more like 20Hz at the appx. 10 watts of audio being delivered to the screen. This was for a 500VDC supply and a screen voltage swinging from 250 to 750 volts. So this is even a little better than the modulator's response at full power (a carrier of 1200W fully modulated). The handbooks say the screen voltage swing need not be as much as the plate, only about 40%. In this circuit the modulator has to supply more power than what goes to the screen. The circuit loaded the modulator with a 58K Ohm load in parallel with the load of the final. The bulk of the AC voltage is dropped from the modulated B+ swing of 5200V in a 50K resistor, which dissipated about 50 watts.

A lot of overkill here in the explanation of this experiment but my point is that it looks like it is better to use a screen modulation winding or even couple in a separate transformer if at all possible. The size of this xmtr is gross, but this applies to anything, scale up and down.

[WA6PBJ]

All very interesting. Screen follows the plate for linear AM plate modulation. Use a choke is one way.

But me thinks: What about the screen bypass caps? These it seems to me would take straight-away the modulated RF on the screen to GND. I sense not so. THe value is for RF...not audio?? Doh!! 

[Steve - K4HX]

More on the screen bypass caps here.

http://www.amwindow.org/tech/htm/scrnmod.htm

[Opcom]

The caps should bypass RF and should not bypass audio too much or there may be an excessive phase shift and the highs may also be attenuated because of the low pass filter created with the choke. Also observe the resonant frequency of the choke and the bypass capacitors that it is away from the audio bandwidth. None of it is too critical.

I use a choke instead of the above schematic because that experiment turned out to have a great deal of phase shift after it was investigated thoroughly. It was an attempt to add a 'virtual' screen winding to the modulation transformer.

In the end the original choke that came with it still works best on this transmitter.

[DMOD]

I think there must be a typo in the last formula unless I missed something in the article.

I think C1 has to be XC1, otherwise, the value for C2 is non-sensical.

Let's assume values for an 813 stage:
Bplus = 2000V, Vsg = 350, Isg = 35e-3, XC1 = 42 ohms for C1 = .001 uF at 3.75 MHz then:


R1 ~ = 79k (rounded up)

R2 ~ = 120k (rounded up)

C2 (with Faudio = 10 kHz) ~ = 0.22 uF (nearest standard value).

The 3 db down point of R1C2 is 9 Hz.


If C2 = 0.6*(1/(twopi*10e3*C1)), then C2 =  9.6 kiloFarads and the

3 db down point would be 2.1e-10 Hz.


Phil

 

  

[wa1knx]

Phil,
        Your right, C1 and C2 should more or less the same. The idea was to reduce the effect of Xc1, with Xc2. I think Don K4kyv published it in the AM exchange a long time ago.  Look at the date of the Derbs original post..

Dean