|[AMRadio] Johnson Valiant Scope Takeoff Circuit Problem|
w0ng at comcast.net
w0ng at comcast.net
Fri Nov 30 18:55:02 EST 2007
I'd like to ask the group, especially those familiar with the Johnson Valiant transmitter, to comment on my phase shift and/or distortion issue. I spoke to John, WA5BXO, about it and pasted his comments at the bottom.
Opinions, advice, comments of any kind would be welcome !!! Here's the issue..........
After installing the new audio driver xfmr I decided to make a small mod I found in the 1964 ARRL handbook showing an oscilloscope takeoff circuit I can add to my Valiant to display a trapezoid o'scope pattern.
I fed the mic input with an attenuated 800hz audio sine wave from an HP audio oscillator and upped the impedance from 600 ohms to HiZ using a Heil mini-transformer.
According to the text in the handbook, I think I'm looking pretty linear at 30-50% modulation but have distortion of some sort at above 50% modulation. At 100% modulation, the upper and lower lines of the trap pattern on my 60Mhz scope each start to form an elongated elipse instead of being sharp, straight lines.
Some of what youre seeing is phase trouble in your monitoring system. There should be only one sloped line at the top and bottom not two lines crisscrossing. The phase difference between the audio that is modulating the rig (at the output of the modulation XFMR) is different than the phase of the audio that is being fed to the scope by just a few degrees.
If you are indeed taking the audio from the top of the modulation XFMR secondary, then the reduction network that you or using has a phase shift in it. If you are sampling the audio from some point in the speech amplifier or driver then there are a lot of places for phase shifts including the driver XFMR and the modulation XFMR.
The best way to feed the scope with audio is to make a resistive voltage divider right off from the modulation point or the bottom of the finals RF choke where there is only B+ and audio, no RF. You will want to make the network from carbon or glass film resistors not wire wound power resistor because wire wound power resistors will have a phase shift at the higher audio frequencies noticeably starting at about 1000 HZ and getting worse as the audio goes up in frequency.
You will need to use a lot of 2 Watt carbons to get the resistance and power requirements needed. The total resistance should be about 50 K ohms or less and the bottom resistor that will shunt the scopes horizontal input needs to be low enough to reduce the audio to a level that the scope needs.
Of course the plate DC voltage should be blocked from the resistor network with a 1000 volt capacitor larger than 0.1uf to make sure that you get good audio coupling at the lower audio frequencies. The RF should be taken with a loop near the low Z end of the tank coil or a separate antenna with a RF choke across the vertical input of the scope.
You dont want anything but RF getting in the vertical input. You will be surprised at how well the separate antenna works especially if you have a little tuner on it.
Back to the modulation linearity:
There does seem to be a slight premature drop in the RF output level on the negative end just before 100% but it is difficult to tell as it may be just a combination of phase distortions between the RF and the audio as described above.
Modulation linearity trouble means that the RF output is not following the voltage that is modulating it. Typically this is only at the extreme positive or the extreme negative of the modulation cycle.
When you have a good network for trapezoid pattern then the sloped edges of the trap pattern should be straight all the way down to the 100% negative point and at least past the 100% positive point. Typically if the pattern doesnt continually go upward at the top but tends to flatten out then you either have too little drive or in the case of tetrodes/pentode type tubes, not enough audio on the screens.
If it goes upward more than it should then you have too much audio on the screens. Similarly an incorrect amount of audio on the screen will cause the distortion near the 100% negative point as well. The negative end is where I think I see some distortion.
Typically, audio for the screens is passed through a screen dropping resistor from the plate circuit. This resistor is used to reduce the B+ that is fed to the screens of the final but also reduces the audio by the same amount.
Here or some examples: Lets say that your or running a pair of 6146s with a plate supply of about 500 Volts and you need 150 volts for the screens grids. The screen dropping resistor does this when the rig is tuned up properly. On audio peaks at 100% modulation the plate voltage will swing from 0 to 1000 and the screen voltage will swing from 0 to 300. Both are modulated 100%, or so you would think.
The screens bypass capacitors, necessary for RF bypassing; will bypass some audio, and more on higher frequencies. This means that the screens will not get as much audio voltage at higher audio frequencies.
This would not be too bad because you really dont want the screen to have that much audio anyway as it causes the distortion that we described earlier where the sloops of the TRAP are not straight lines, Except that it is not the same for the lower frequencies and there is a phase shift caused by the screen bypass capacitor reactance.
Typically the audio on the screens should hit about 40% modulation when the plate is at 100%. The thing we need is to reduce the audio voltage that gets to the screens with out reducing the DC on the screens. This can be done in a number of different ways.
One way would be to have a special screen winding on the modulation XFMR but that is a pretty extreme modification. An easy way is to add an audio only load from the screens to ground. The resistor should be about ½ the value of the screen dropping resistor (some experimentation is necessary) and power rating to not get too hot while modulating.
After it will be absorbing some audio as you modulate but only a few watts of audio if that much. To avoid phase shifts the resistor or resistors should be carbon or glass film not WW. There should be a capacitor in series with it to block the DC, after all we dont want to change the DC but we do want to load the audio down some.
The capacitor should have a rating larger than the maximum DC to the final (not just the screen voltage) and will need to be about 1 5 uf (enough to pass all the audio frequencies).
Before experimenting with screen audio voltage reduction, you really need to work on the phase shift in the monitoring networks to get accurate results first.
John Coleman, WA5BXO
More information about the AMRadio mailing list
This page last updated 13 Dec 2017.