I thank all of the contributors for their replies. I have looked over the old transmitter and it seems to have a big dropping resistor to the screen. It gets fed from the HV supply, not from the modulation transformer. I have looked for a choke and do not see one. There is a clamper circuit with 2 tubes to short the screen voltage to ground if there is no drive. So a choke could be in there that I missed. But so far I have not seen a choke. But I'm not sure I know what kind of choke to look for. Not an RF choke, or a power supply choke. What shape would it be?
In your system you really don't need a choke specifically. A resistance will work fine if the resistance is high enough and if the screen is not bypassed to ground for audio. What you really need on the screen is a high impedance back to the HV, most likely 5,000 ohms or more would be enough, so the screen can "wiggle around" at an audio rate.
A choke is required when the screen dropping resistance can't be made large enough. The choke would need a reactance of several thousand ohms at the lowest audio frequencies.
It would be much better if the screen was fed after the modulation transformer, so the modulation transformer modulates the screen supply voltage along with the plate. There really is a very specific percentage of modulation that must be applied to the screen and plate to make the modulation linear.
Look at what my Globe Scout has and what I did:http://www.w8ji.com/Heising%20modulation.htm
SO the screen seems to get its voltage from the wrong place. It makes me wonder if this old transmitter was ever a working unit. Another telling sign is that the shielded audio wires to the audio gain on the front panel went by a power transformer and it picked up a bad hum there until I re-routed the cable. That made me wonder how it could have been operated with all that hum.
When I got the unit I thought that I was bringing a good transmitter back to life. Now I think I have a poorly designed one that needs re-engineering.
Most old gear is that way. Even some commercial rigs are full of small technical problems.
I have used a monitor to listen to the signal on AM. It sounds a little distorted with the highs cut off slightly. But the audio is very understandable.
That's all most people cared about.
Looking at it with a monitor scope the envelope looks pretty good up to about 50% modulation. After that it shows distortion. The curve goes up to a peak OK but has a flat line on the downside for 1/3 of the way down. And the same on the bottom it is a mirror image of the upper curve. The null at the center Shows a heavy line in the shape of < . This is just to the right of the node, just where the next positive rise begins. I do not know how to interpret these lines.
Too bad you can't post pictures. Even if you could though, there probably is not enough info to tell what the problem is.
One thing I have been thinking about is The loading of the transmitter. I am running it at about 2200 volts and 290 ma. I think this is set up should be loaded heavier but I don't want to strain the power supply further. So perhaps I should change the taps on the modulation transformer to adjust to this load. Being homebrew I do not know what the design intended or if the parts ever worked together well.
They almost always modulate BETTER at lower power, not worse. I am starting an AM modulation specific web page. Eventually there will be enough on it to help, but it might be a year to complete. http://www.w8ji.com/amplitude_modulation.htm