reverse screen current!

[N2DTS]

I built a quad of 4x150 tubes for screen modulation, it worked great with a solid state screen modulator.
It was good for 300 watts out with little air on the tubes, that was not running it hard.
I took the tubes out and tried using them in another RF deck, but because of an oversight I discovered when I took things apart, it did not work out.
I put the 4 tubes back in the original rf deck and now have reverse screen current.
If I increase the screen voltage past a low point, the screens seem to generate their own voltage and the plate current goes up a lot.
I have negative screen current, can turn the screen voltage all the way off and still have voltage on the screens, and power output (on frequency) and can kill it with removing the drive or plate voltage.

Its not a parasitic, its right on frequency, and goes away with drive.
I did not change the RF deck, I swapped out one tube at a time and had no change, I added more bypassing, increased the grid bias, it always does it.

Reading the care and feeding of power grid tubes, they talk about this happening, the cure is a nice stiff bleeder resistor on the screen supply, which you can not do with screen modulation.
Why it started is unknown, I was running the tubes at slightly less then the 6 volts recommended.

I may try changing the rf deck to a quad of 4-125's.

The quad of 4x150's still work well with plate modulation, I just wonder why the reverse screen current went way up on these new tubes when I tried them in something else and then put them back in the original rig. They were not overloaded....
Odd stuff!

[w4bfs]

Eimac's later versions of 'care and feeding ...'  had more discussion about tetrode neg screen current as a function of secondary emission ... the external anode tubes have a copper plate (internal coating ?) and seem to exhibit more of this .... the screen 'dummy load' value is such that it draws 10% of plate current at nominal screen voltage .... if you don't have this a regenerative condition can arise with the screen current and lead to screen overdissipation ... everything about the negative screen current is loading dependent ....tricky

[WD8BIL]

Does the screen supply in the "other rig" go on and off with the plate supply?

You didn't leave screens on with plate off, did you?

[W2VW]

Sounds like the plate loading changed.

[N2DTS]

Its the same rf control deck, I just made a new plate up and replaced a pair of 813's with the quad of 4x150's.
So same control deck, power supply, just a different basic rf deck.

I tore the quad of 4x150's apart and am putting in a quad of 4-125's.
The 4-125/250/400's are supposed to work well with screen modulation.
Four tubes should do 300 watts of carrier.
And no reverse screen current.

I have a pair of 4x150's with a tube screen modulator that works very well.
The 4x150 tubes always had some reverse screen current, and a 100K ohm resistor was enough to take care of things.

[WB3JOK]

Any tetrode can exhibit reverse screen current, usually when loading is insufficient... that is why many amp designs, including mine, use a shunt regulator for the screen supply (so that the regulator can absorb the extra screen current instead of "running away" as when fed from a dropping resistor).

[N2DTS]

I think the 4x150's are worse then other tubes because of the oxide cathode and close spacing on the little tubes.
Not a problem with normal operation, but a problem with some screen modulation setups maybe.

[w4bfs]

it will be interesting to see how the quad 4-125 rig does .... how do you plan to equalize the tubes for gain variations so that they share the work equaly ?

[N8ETQ]

     I have just taken to putting the screen bleeder right in the RF deck.
Right at the socket(s). This props the Iscr meter up 10 ma or so and if it
goes negative, well, it goes negative. It's a loading issue. Never was a
problem for me. I guess I don't know what you mean by "a solid state screen modulator."

    All the 4X150,250 class tubes really need a regulated screen supply.
There pretty "Dirty" without it. You can have screen voltage applied
without plate voltage provided the tube is biased off hard.

    It's a little vague but I would say from your original problem
description you may have killed your bias supply. Easy to do with that
big grid connection on the tube.

GL

/Dan

[N2DTS]

The solid state screen modulator is a modulator for the screen supply, puts audio on the carrier...
The bias supply is fine, and I can get up to -500 volts bias.
I have fixed (adjustable) and grid leak (adjustable) on this setup, a variac bias supply and a big wirewound pot for grid leak.

The RF deck has vernier drives on everything and a chart to enable things to be reset very close on band changes, so the loading was not changed, plus I tried different settings of everything, bias, grid leak, drive levels, plate voltage, screen voltage, screen modulator in or out, even different pie net coils and neutralization settings.

[N8ETQ]

  Well, Something had to change, I assume you are driving the
screen directly with a solid state device and not a xfmr.

/Dan

I put the 4 tubes back in the original rf deck and now have reverse screen current.

[N2DTS]

I did not change the RF deck other then to pull the tubes out of it and put them back in.
It tuned up normally, same dip at the same points, same grid drive.

It would have made a good plate modulated rf deck, put out plenty of power, 1000 watts carrier, but I do not need that and wanted to do more screen modulation experiments.

As far as equalizing the load through the tubes, very little is needed in most cases, a resistor in each screen and plate lead helps balance things it seems.
I will be able to tell more by plate color with the 4-125 tubes.

[WB3JOK]

That's what I did in my quad 4-125A amp... a grid bias adjustment pot for each tube, initially turned all the way "off". Apply plate and screen voltage, set the 1st pot so its tube draws 25 ma cathode current, adjust the 2nd pot until the meter reads 50 ma, 3rd to 75 ma and of course the 4th to 100 ma 

I do make fine adjustments thereafter by equalizing the color of the four plates.  If I had an infrared pyrometer I could probably set the current matching very closely

[N2DTS]

Well, I got it done and ran some tests.
Four 4-125's (what I have) in place of the four 4x150a's.
2500 volts on the plates, about 200 ma, 150 watts out, sounds good, no strange or erratic action, all the plates glow evenly.
Lots of heat at 200 watts carrier and the plates are showing a dull orange, about 4x pep, no carrier shift, everything seems to be working well, but not a great level of power output.

I have a 100 ohm wire wound resistor in each screen and that seems to balance things well enough.

I need to get some 4-400's!

[AB2EZ]

Normally (but not in this case), you would expect a fraction of the electrons flowing from the cathode to the plate to be "captured" by the screen as they pass through the screen. This would lead to normal (positive) screen current. The electrons accumulating on the screen would flow out of the screen, via the screen pin, to ground, via the screen supply... implying a positive screen current.

However, in these tubes, you are dealing with the following phenomenon:

Electrons passing through the screen can collide with the screen to produce "secondary" electrons. These collisions can result in (on average) more than 1 secondary electron being emitted from the screen, for each incoming (primary) electron that collides with the screen. These emitted secondary electrons continue on to the plate (they are not recaptured by the screen) because the plate voltage is higher than the screen voltage.

Therefore, the screen (losing electrons) becomes more and more positive with respect to ground. Electrons will flow from the screen supply into the screen, via the screen pin, to replenish the secondary electrons that are being emitted by the screen.

This flow of electrons into the screen, from the screen supply, is "negative" screen current.

The average number of emitted secondary electrons that leave the screen... for each incoming electron that collides with the screen, will be larger if the voltage between the screen and the cathode is larger (more energetic collisions). This results in a positive feedback phenomenon.

I.e. as secondary electrons leave the screen (traveling to the plate), the screen becomes more positive... and the average number of secondary electrons per collision increases.

Reducing the DC resistance of the external path between the screen pin and the cathode pin will reduce the voltage on the screen ... and moderate or eliminate this phenomenon.

Since the onset of the emission of secondary electrons from the screen, and the average number of emitted secondary electrons per colliding primary electron, is very sensitive to the screen-to-cathode voltage (and therefore the energy, e x V_screen, acquired by the primary electrons when they travel from cathode to screen)... it is possible that one will not see negative screen current until the screen-to-cathode power supply voltage is increased to the point when the production of secondary electrons (on average, more than 1 emitted secondary electron per incoming primary electron) begins. At that point, the positive feedback mechanism (described above) will take over.

Note: This process (secondary electrons produced by collisions between primary electrons and the atoms of a material) is employed in photomultiplier tubes.

Stu  

  

[N2DTS]

Glad you are back Stu!
The 4x150's really do that a LOT.
Something changed in the tubes when I tried running them in something else.
Originally, with new tubes, I had to swamp the output of the screen modulator with a 100K resistor, and still could not run the deck past about 300 watts carrier without it taking off.
After using the tubes in something else, maybe with the filament voltage out of spec, I could only get the screen voltage up to about 100 watts carrier level (100 volts?, I did not note it) before it took off.
A modulation peak could trigger it.
The screen modulator circuit looks like an open circuit to the reverse screen current, so it did not work well.
The oxide coating of the cathode likely coats the screen (and everything else in the tube) and causes the reverse screen current problem, along with the close spacing/high power density.


I could not think of any way to correct the issue, so I put 4-125's in which do not have the problem at all.
Lower power output and not as clean as the 4x150's maybe, but stable up to max rated power output.

A transformer driving the screens off a stiff screen supply might work, I should try that again.
 I wonder what transformer would work best on the screens (of four 4x150's).
I should try an 8 ohm single ended output transformer and see what happens.
I tried that before but had no clue about how to adjust the rf deck for screen modulation, it takes a feel for the process I think...

[WB3JOK]

The screen modulator circuit looks like an open circuit to the reverse screen current, so it did not work well.

I could not think of any way to correct the issue, so I put 4-125's in which do not have the problem at all.

How about a shunt regulated screen supply, which can absorb the reverse screen current...

[N2DTS]

Its not a power supply issue, the screen modulator is between the power supply/bleeder resistor and the screens.

[N2DTS]

ckt

[WB3JOK]

I see... well, then the only simple way that comes to mind is a "swamping" resistor from the screens to ground. Wastes a little modulator power, but better than screen runaway from your 4X150's...

[N2DTS]

I would have to go so low in resistance as to make the modulator work too hard.
Better to just use the 4-125's or a pair of 4-400's.
Could do six 4-125's....in a circle...