S-meter/AGC activation by strong broadcast stations outside the ham bands

I have hopefully solved the problem by inserting a bandpass filter after U2B the AGC IF amplifier and before the AGC rectifier.
The filter is centered at 153kHz and is 20kHz wide It does not interfere with normal AGC action as far as I can tell. I do not hear any plopping on SSB or CW signals. All it does is "cleaning" up the AGC IF signal at 153kHz.

Why BP filter the AGC IF signal?

Out of band broadcast signals leaking into the IF-amp past the narrow crystal filter will produce IF signals several hundred kHz away from the normal 4.915MHz IF freq. This should not happen, but it does when I use the OP1 filter on the SSB board.
Out of band signals produce AGC IF signals far away from 153kHz and also produce AGC voltages and S-meter activations which are unwanted. This should not happen!

The 4.915MHz IF signal is mixed with the 5.068MHz AGC-LO producing the 153kHz AGC IF signal which is then amplified by U2B and rectified to make the AGC voltage. U2B is a wideband audio amp and untuned.
Any rf signal leaking into the wideband AGC IF amplifier U2B LM833, (U2 has a bandwidth of more than 10MHz), will produce AGC voltage and S-meter indication, inserting the BP filter will attenuate signals except around 153kHz.
Signals most likely to leak into the AGC IF amplifier are the 5.068MHz LO and the 4.915MHz IF signals, if the balance of the NE602 mixer is not perfect.

Signal leakage in OP1 on SSB filter/board

But there is a leakage in my K2 when using OP1 filter on the SSB board! Using the CW filter on the RF board does not produce leakage, only the OP1 filter on the SSB module.
The best cure for this problem is to get rid of the leakage in the SSB/OP1 filter on the SSB board, but it is not easy to do, the leakage is probably due to a ground loop pickup, as it is only present when the IF signal is routed into the OP1 filter across the SSB board.

The official solution by Elecraft

Of course you should do the "tightening up the bandpass filters mod" as suggested by Elecraft as it goes a long way to resolve this leakage, as the unwanted signals are weaker.

The 153kHz BP filter


Schematic of the filter

Here is a schema of the filter, it uses standard component values for L and C, the values chosen resonate at 152.5kHz. Using 5% tolerance components is recommended. No tuning is needed.

The 3.3nF and 330uH I used are 5% resp 10% tolerance. The 2.2k resistor controls the Q of the circuit and the 10nF cap blocks the DC from the IF amp. Better use a high Q choke, mine has internal resistance of 5 Ohm and the insertion loss of the filter is about 1dB, according to Spice3.
The center frequency is 152.5kHz, -3db bandwidth is about 20kHz, -20dB points at 100kHz and 350kHz.

Does it work?

I have checked the workings of the new filter with a S9+40dB signal from a small Xtal osc on 3.579MHz in the 80m band, the test was done using OP1 Xtal filter on the SSB board, Preamp On.
See the S-meter response for the above case on the next figure. S9 corresponds to 6 LEDs lit. The red curve is the S-meter response to the signal and the green curve is the S-meter response of the AGC IF mirror image, showing the response of the BP filter, the S-meter would have 5 LEDs lit all the time if not for the filter. The AGC IF freq is 153kHz which makes the AGC IF mirror image appear at 306kHz higher frequency on this band or at 3884kHz.

The filter takes care of the S-meter activation problem, except when the offending signal falls in the 200kHz wide image window which is at 306kHz lower freq on 160m to 17m bands, and at 306kHz higher freq on 15m to 10m bands than the operating frequency.
If the offending signal is more than about 400kHz away from the operating frequency there should be no AGC action from it with the BP filter in.
On the 17m band I usually got readings around S9 from a strong broadcast station around 17.660MHz, now I checked this frequency and there was a S9+40dB (full scale) station there, but on 18.068MHz the S-meter indication was only S2 and on 18.150MHz only S1, so the filter has cured the situation for me, and I do not need to turn off the AGC anymore!!

How I did the mod

I have K2/KPA100 so I removed the small plate on top and could get access to the Control board from there. I unsolder-ed and lifted the top end of R3 the 10k resistor, which is located next to P4 the flat cable connector on top of the Control board ( on the upper left side when looking at component side of the Control board) and inserted the 2.2k in the hole and soldered the two together and connected the resonant circuit via 10nF the DC blocking cap. I guess there are better ways to do the mod.

Setting the AGC threshold, (tnx Sverre, LA3ZA)

The AGC threshold should be set (readjusted) -R1 on the Control board - to compensate for the attenuation of the filter, which affects the AGC sensitivity. The filter will reduce gain in the AGC loop slightly, meaning that the overall gain of the K2 will increase. This can be compensated by adjusting R1 on the control board so that pin 5 on U2 reads about 4.1 Volts (Elecraft's serial # 3000+ recommended value is 3.90 Volts)
I needed to change R1 from 51k to 100k to get 4.1 Volts on pin 5, my K2 is #2024.

Listen to the AGC IF signal

If you have an AM LW radio try listen to the AGC IF signal on 153kHz, just hold the radio close to the upper left side of K2, makes it simple to listen to AM stations using the K2, does not it!!
73
Matti, TF3MA
----

A better filter?

A better way to do this is to insert a narrow (abt 50kHz) BP filter ahead of the AGC mixer, this requires making a 4.915MHz tuned circuit with a loaded Q of abt 100 which is not easy, but could be done using 1uH+1000pF across pins 1 & 2 of U1, the mixer and a 1.5k resistor in series with C181 the IF out DC blocking cap on the RF-board. The 1uH inductor must be tunable +- 10%.
This filter has insertion loss of abt 5dB and 3dB BW of 200kHz and -20dB BW of 1200kHz. Not very impressive.

Last modified: Sun Jan 26 17:39:18 GMT 2003