6BA6 -adjacent signal causing quieting on HF modes

[mag_therm]

This effect becomes obvious when my vintage Lafayette HE30 is sending audio to sound card and hence waterfall.
In the old days, the effect would have not been so apparent as only audible, and I never noticed it.
With AGC off the 6BA6 as first RF stage only has grid leak bias. **
When a strong signal ( maybe around S9, -73dBm , 50 uV) about 10 to 30 kHz away comes through,
 for example strong CW or BPSK31 when receiving FT8,  the receiver gain reduces as the strong adjacent signal increases.
The strong signal does not break through, so it is not heard, and the waterfall fades toward dark blue, as the audio level of the wanted signal reduces.

I doubt that 50 uV would vary the bias hence gain of the 6BA6 and adjacent interfering signals are way below IMD levels.

I did not find reference to this effect in some search, or my old reference book.

Any insight from the tube ones would be appreciated. Thanks

** has cathode bias 330 Ohm, 10.1 mA, 3.35V

[RoV]

Are you sure the cause is not in the following IF stages? 10 to 30 kHz offset will pass several of them. The level will be higher there.

[mag_therm]

Yes, it could be. I will open it up, check all biases, and try to find the cause of the variation.
Tubes have been running almost continuously for nearly 3 years. Maybe just a tube  needs to be replaced.

[Wallace Gasiewicz]

What you are describing is "desensitization" and caused by nearby frequency strong signals.
It is kind of "Normal"
I do not think you can do much about that with your current receiver.
Some radios had a tunable "rejection" circuit to help with this.
It is a well known problem more so with Old Receivers.

[mag_therm]

Hi Wally,
Thanks for introducing me to the correct term, "desensitization". There is not much on web, but this paper
by Gavan and Shulman dated 1984 has a good description.
https://www.researchgate.net/publication/3156078_Effects_of_Densensitization_on_Mobile_Radio_System_Performance_Part_I_Qualitative_Analysis

When receiving digital modes, I presently have the AGC off, selecting MVC . However in this receiver, that mode puts the 1st RF amp
into high gain, with a low bias of -3.3V on the 1st RF.
Then the IF gain is manually adjusted and is about 15 dB turn-down to get the suitable level into sound card.

So the first  try will be to increase bias on the 1stRF 6BA6 to about -7 volt which both reduces the gm, and allows
more linear range ( visually on the curves)  with  +/- 7V of swing between cutoff and positive grid.

The above paper mentions that the tetrode characterstics of tubes and mosfets make them less sensitive to desensitization.

[Wallace Gasiewicz]

There is quite a bit about "Blocking Dynamic Range" in my old ARRL Extra study guide.

[mag_therm]

Thanks, Yes, the current Extra Manual has  a paragraph and questions about "Blocking"  at section 7.16, but only the effects, not the mechanism.

Today I have reduced the 1st RF gain, by increasing the cathode resistor from 330 Ohm (-3.3V ) to 1000 Ohm ( -7V)
In doing that, the cathode bypass capacitor has been removed, to add feedback, and the gain is now about 50 if the load line is about 50000 ohm.
That causes more gain to be needed in IF, and there is still enough IF gain to bring the noise to 0dB FS on the sound card.

I was simulating the nearby strong signal with the sig gen at 4 kHz above passband. However that does not fully replicate the effect
of actual strong signals.
After the mods, the sig gen level was 65 uV ( -71dBm ) to desensitise incoming BSPK31 at 4 uV to 10 uV.

I will try it tomorrow when ft8 is strong.

[T3sl4co1l]

Not sure if RDH4 covers it (my copy isn't OCR'd unfortunately), but I see a reference in RDH3:
https://pearl-hifi.com/06_Lit_Archive/02_PEARL_Arch/Vol_16/Sec_51/4394_Radiotron_Designers_Handbook_3rd_Ed.pdf
page 242 (234)

Tim

[szoftveres]

Just looked at the schematics and wondering if the 'AGC off' switch is working properly in your radio - the IF gain control is controlling the IF stages independently from the RF 6BA6, which is either controlled by the AGC or the AGC is grounded by the AGC off switch.
If this switch doesn't close properly then you'll still see AGC action.

[mag_therm]

Hi Tim:
Thanks for the RDH.
Just above the RDH "Blocking Interference" test is a note that it is not necessary to test with the interfering signal closer than 10 kHz, as the minimum channel spacing at that time was 10 khz.
However the spacing presently on 20 m band for example has the bpsk31 band from 14.070 to 14.073, and the ft8 band only 1kHz above from 14.074 to 14.077.
So correspondingly tougher for a purely analog receiver now.

Hi szoftveres,
Yes, the MVC switch position is pulling the 1stRF grid resistor hard to ground, and the AGC, when selected is rectified from the final IF filter which is +/- 2kHz wide.

My understanding of the desensitization or blocking is now as follows, and it may not be fully correct.

When the interfering signal "wi"  mixes with the wanted signal "wf" on the non linear surface representing the anode current as functions of the anode and grid voltage,
there are intermodulation products of 2*wi, 2*wf, and sum and difference wi+wf and wi-wf.
the original frequencies wi and wf remain, but are both reduced in amplitude as part of energy balance before and after the intermodulation.

This effect probably occurs in all RF and IF gain stages, which might be why the RDH manual shows testing on the complete receiver.
As to why the atmospheric noise is also uniformly attenuated by "wi" ( shown by waterfall bands of dark blue all across the band ) I do not know, unless it is due to the same mechanism..

Today's testing after mod to the 1stRF stage yesterday by both reducing its gain, and giving it negative feedback, is showing good results, I think.

It is not possible to quantify due to to varying signal levels, but I can see that the desensitization is reduced, while not eliminated.
This scrot shows two darker bands across the bpsk31 waterfall caused by ft8 stong signals.
https://app.box.com/s/shxv12tthl927n9ocihf27ejbd9j9f28

This one was caught by chance, showing a strong cw test signal  in band . Note how the waterfall is showing a band of lower level on both the bpsk31 signal and the background noise.
https://app.box.com/s/wyrprtloc7w9z7z9ug3tda4hspovdxpa