20m bandpass filter modification for IC-706MkII

Here are some hints and tips for adding a 20m band pass filter for the Icom IC-706MkII(G). Sorry, this is not meant to be a 'step-by-step' description on how to do it, because components available / required can differ - my intention was just to provide some indications on basis on my experience....

Needless to say, the IC-706 service manual (download as pdf - 13MB) and a RF Spectrum Analyser are a 'must have'  to perform this mod under adequate conditions.

Why this modification?

When I purchased this nice little rig, I immediately realized that the reception on 20m band was very poor during evening hours, because of strong (S5 to 7) intermods and spurious sigs all over the band....

As antenna I was using just a simple quarterwave GPA.... This was really bad, because 20m is the most band interesting during evening hours ... and weak signals were not audible (especially CW) under these conditions. This was observed with the built-in PREAMP off (switching it on, intermods were reaching S9+), while switching on the 20dB ATT had indeed a positive effect for intermods, but really kills out the weak sigs as well...

It became clear that this rig really suffers from intermods due to wideband input circuit (for 20m band : 8 to 15 Mhz preselect) and strong (S9+60) broadcast stations in the 12 and 15MHz adjacent bands.

Now OK, I understand that it is quite a challenge for equipment designers to integrate front-end filters in such small package, but I can tell you that my (even more compact!) FT-817 under exactly same conditions performed much better ;o) ! I later found out (ARRL test result) that the 3rd order IP of this rig is only about -2dBm ... confirming that it is absolutely not able to handle strong signals.

Most recent middle-class transceivers are nowadays in the region of +20dBm, or +30 to +35 dBm for the high end sets (indeed costing +10.000$).

Further findings around this trouble :

Maybe 'bad luck' with my unit ? Well I tested performance with another set .... exactly same ! The owner of this twin never had observed any intermods ... but was always working on 'magnetic loop' antenna's, which are in fact very efficient 'filters' !

The solution ...

I have been thinking of a solution that would meet following requirements :

So, a solution that would offer a perfect RX on 20m while not disturbing anything else .... this lead me into the design of a switchable 20m bandpass filter which would be front-end mounted

Hear the difference !

I recorded 3 audio sample files in mp3 format, where the effect of the filter is demonstrated (it is switched on/off several times). These recordings are just representing average conditions, I really didn't concentrate on looking for extreme conditions. During these recordings, the audio DSP was always OFF. As explained further, when BPF was inline, the PREAMP was switched on as well. When BPF off, the PREAMP was switched off as well (File size download is about 250 kB each)

The modification ...

The bandpass filter:
 A BPF (second order Butterworth) with centre frequency on 14.175 MHz must be constructed, offering steep edges, and acceptable in-band attenuation (3 to 6 dB max).
  • L1 and L2 are T37-6 toroids
  • 2 and 11 turns of small dia copper, enameled( #24 AWG wire)
  • the 2 capacitors are approx 220pF +/- 15 pF. I used fixed ceramic type instead of variable, because of smaller physical size, and by putting the right combination in parallel (working in steps of 4.7 or 2.2 pF), found the proper tuning around 14.175 MHz.
  • the coupling capacitor is 4.7 pF to 10 pF max (check for over-coupling)
  • both 0.1 F capacitors are required for Icom circuit
  • the BPF (together with switching relay) is constructed on a 'sandwich' PCB , size identical to a X-tal filter.
Spectrum center=14MHz, hor scale 500kHz/div, vert scale 10 db/div See actual performances:
  • -4 dB @ center frequency
  • -6db @ band edges 14.000 and 14.350 MHz
  • -20 to 30 dB @ 1 Mhz from band edges
  • -50 to 80 dB @ 3 Mhz
The switching logic:

The filter is switched inline through a subminiature (SMD - only 14mm high) double pole relay of MATSUSHITA I could recover from a scrapped SMD PCB, operating on 5v DC.

 The filter is switched on (relay energized) when

(1) the 8.000 to 15.000 Mhz band is selected and

(2) the RF-preamp is selected.

The 'pass through' circuit should be decoupled by a 100nF capacitor (replacing C109 of original circuit).

I have chosen for this combination, because the RF-preamp would cancel out the attenuation of the BPF, and the ATT function is memorized per band by the IC-706 !

In order to realize the switching, following signals are to be considered:

Consequently, both signals can be summed by diodes to insure the desired switching logic, and then reversed by a small signal transistor (BC547) driving the relay switching transistor (open collector).

Only when both control signals are 'low', the relay is energized and the BPF switched 'inline'. Both signals are immediately available from respective PCB's without the need to remove them....

The logic switching circuit was mounted on a small PCB, this was located near the RX preamp IC (IC151 PC1658G), soldered on the edge of HPF board / 1st Mixer.

The integration in transceiver:

(click on pictures to zoom)
Overall view of the modification
Close-up of the BPF (same size as X-tal filter, mounted with dual side adhesive tape on main PCB). Left & right the toroids, in the center the relay. On top of relay, a SMD LED is mounted as visual indication that BPF is 'inline'. This LED is visible through top cover vent holes.
Close-up of the signal pick-off at FILTER UNIT :
  • RL8 at junction of C30 and C93
  • RF signal : remove C109 and tap off from each side
Close-up of switching logic PCB. This small PCB is soldered just next to HPF casing. Components (mostly SMD) are soldered directly on PCB (used as support).
Close-up of +5v DC tap-off (from UT-102 connector on main PCB)

Using the rig with the MOD ....

Normally, the PREAMP is always used on the 20m band and consequently the BPF in use. When switching to other bands, the PREAMP setting is memorized.

The signal difference on the S-meter with PREAMP+BPF switched ON and OFF is about 1/2 'S' point, however the signals are significantly 'cleaner'.

Should you consider to carry out this mod, I wish you good luck ! As alternative, consider to use an external 'filter' : for example, intermods are already reduced when using an antenna tuner.... or completely disappearing when using a 'magnetic loop' or other narrow band antenna!