IC-R7000 - better SSB selectivity

[N8YX]

Anyone who has owned various Icom or Kenwood rigs which utilize a 455KHz IF may have encountered one of these beasties if the rig's covers were ever removed for servicing or upgrades:

4948

This particular filter is a 6-pole unit and is spec'd for a 6dB B/W point of ~2.4KHz. It's also rather small, which comes in handy for what follows.

The IC-R7000 receiver uses a 455KHz IF strip, and employs three Murata 4-pole filters on the IF/Detector Board: One with ~15KHz/6dB B/W, one of 6KHz/6dB B/W and one of 3KHz/6dB B/W. The latter (F16 on the IF Board schematic) is used for SSB reception:

4949

Good, but not optimal. SSB sounds a bit "hissy" and if you use the receiver on a crowded band (say, 10M during a 'phone contest) stations adjacent to the desired frequency can create a noticeable amount of interference.

I freed up a couple of K5s recently when I installed FL-44As in two of my R-71's accessory SSB filter spots. The '44A is also a 2.4KHz B/W unit but its shape factor is a good deal better than the K5...which (of course) is a good deal better than the plastic cube unit. At least, on paper. It is also shielded, whereas the OEM unit is not.

If one is willing to get creative with a solder sucker and a Dremel tool that has a very small dental burr installed, one can remove the OEM filter and modify the IF Board so it will accept the K5 or another K-series filter of identical form factor. There is just enough room to accommodate the larger filter. A couple of components must be desoldered and relocated to the foil side of the board, where they are reattached.

Pics of the finished assembly to follow, and when I rework the next IF Board I'll take "before" and "in-process" photos.

I might try to locate a narrower K-series filter. While an improvement as far as the R-7000 is concerned, the K5 is not the equivalent of its larger 8-pole cousin. But there's not enough room to mount one of those, either.

[N8YX]

Details of the 'surgery'. You don't have to be precise with the Dremel, but a steady hand helps - and there's not a lot of room in which to work.

First, the top cover has to come off the R7000 then the leads going to the IF Unit are removed. (There's one wire bundle which is soldered to the left-hand side of the IF Unit; no need to disconnect the far end from the RF Unit. The bundle can be left in place for this modification.) Next, remove the screws which hold the IF Unit down and gently free it from the parent compartment.

A look at the IF Unit and the area which will be worked on. F16 (topmost filter), C45, R48 and D12 must be removed in order to make room for the larger filter. The cap, diode and resistor are part of the filter switching circuit and will be moved to the bottom of the board.

4984

Underside of the same board prior to desoldering:

4985

[N8YX]

Area of IF Unit under the now-vacant F16 spot. Sharp viewers will notice that I removed two filters (F15 and F16) - I'll explain this later on. If you're just interested in improving SSB performance, however, only F16 and its previously mentioned switching components need to be removed.

4986

Top of the IF Unit. The filters, C45, R48 and D12 are set aside for later reinstallation.

4987

[N8YX]

Get your Dremel and a couple dental burrs out and let's make some holes. The filter input/output terminals are towards the top of the photo, while the signal and case grounds are at the bottom. The input foil's lead hole is enlarged, a corresponding pair of holes for ground and output are drilled near D8 and L6 and a spot for the filter case grounding tab is opened up near R54. We're going to use that partially drilled foil strip as a new ground area and bridge it to the existing ground plane.

4988

I've gotten a bit ahead of myself here with the photos - the 'output' terminal of the filter (lead nearest D8 ) needs to be isolated from the ground plane. Using the smallest dental burr you can find, carefully make a "U"-shaped cut which results in an island approximately 3/16" x 1/4" or so around the filter output terminal. Ensure this foil is completely isolated from the ground plane then scrape and tin it and the nearby hole for filter signal ground...then install filter and solder into place.

Bottom of board with "island":

4989

You can also see the filter's case ground tab soldered into its new grounding spot.

[N8YX]

Top view of IF Unit with filter installed:

4990

It just fits.

Now reinstall C45, R48 and D12 on the bottom of the IF Unit, paying attention to positioning of D12's cathode (same spot it was originally routed to) and R48 - underside of board, connected to the same 3-hole foil which it was originally inserted into. C45 runs from that foil trace to any nearby ground.

4991

The anode of D12 and the end of R48 route to the little "island" where the filter output terminal is soldered. These components switch the filter output in and out of the IF circuit.

[N8YX]

Check your work and install the IF Unit back into its compartment, then connect the associated wiring.

Put your newly modified R7000 on the bench and warm the set up for an hour or so to stabilize its temperature. Then adjust L15 (USB BFO set) and C134 (LSB BFO set) per the service manual.

[K7SGJ]

Good post. It never ceases to amaze me how ingenious hams can be. A side effect from terminal cheapness, I would guess.

When it comes to enlargening holes on a PCB, or making new ones, I use these if I can't or don't want to use the Dremel. They are very easy to control and they help prevent damage that might be done with a power tool. They came in large variety of sizes, and they are also good for making the right sized hole should you want to swage in a feed through eyelet.
4992

I have a bunch of different sizes (and spares) and they are perfect for electronic work as well as clock making. There is another style that is like a finger pin vice that allows you to use standard bits from about an 80 up to 1/16 or so.

[N8YX]

Some post-modification notes:

1) The CFJ-series filters will occasionally drift in frequency over time. If one looks at the Murata datasheets, their curves look fairly linear with regards to center frequency and high/low cut points. In the case of a 'drifted' unit, one sideband will appear to be a lot "wider" through the filter than the other and it'll be difficult to equalize the tonal quality between the two. The filter curve in one of my modded units favors LSB, another favors USB and the remaining two are normal. Go figure. I may try to realign the entire oscillator chains in those receivers some time and see if that helps; doing so may move the receivers' reference frequencies closer to the proper SSB 'center'.

2) CFJ455-series filters are made in a number of bandwidths: 'A' is widest (around 20KHz/6dB), all the way through 'L' (1KHz/6dB) and beyond. One could pick and choose the bandwidths they want - and even replace all three 4-pole units with better quality parts. This may be useful in a receiver which experiences excessive adjacent-channel interference. In the end, it's a question of cost vs. performance and the casual user would probably balk at an additional $100 cost for features he or she may never use.

3) About those little plastic filters - they also come in a variety of bandwidths. The stock AM filter in the R7000 is a 6KHz/6dB unit and is way too wide for an RF-polluted area. In two of my receivers I swapped them (E-series) for the K-series SSB units that were removed for the upgrade. A bit too tight for some folk's tastes but this is a definite improvement in a noisy environment. I'm on the hunt for F/G/H/I/J-series 4-pole units and will try each in the AM filter spot as they are found.

[N8YX]

Good post. It never ceases to amaze me how ingenious hams can be. A side effect from terminal cheapness, I would guess.

When it comes to enlargening holes on a PCB, or making new ones, I use these if I can't or don't want to use the Dremel. They are very easy to control and they help prevent damage that might be done with a power tool. They came in large variety of sizes, and they are also good for making the right sized hole should you want to swage in a feed through eyelet.
4992

I have a bunch of different sizes (and spares) and they are perfect for electronic work as well as clock making. There is another style that is like a finger pin vice that allows you to use standard bits from about an 80 up to 1/16 or so.

I have a pin vise and small bits too, but darned if I can find the thing at the moment!

[N8YX]

I may try to realign the entire oscillator chains in those receivers some time and see if that helps; doing so may move the receivers' reference frequencies closer to the proper SSB 'center'.

There's definitely an echo in here.

Got receiver #4 modded this morning...and try as I might, I could NOT net that thing on SSB to save my backside. Out came the counter and the probes...got the BFO and 3rd LO set per the manual and found the tuning to be a consistent 1.5KHz off.

Hmm.

If anyone's ever had a -7000 on the bench...suffice it to say they're WELL shielded and as big a PITA to get apart as the old Dak Mark IX/X CB sets. Both of the LO test points are easily accessible if you have the correct mini-plug/coax combo. Of course, they were off frequency. :wall:

After the radio's inner covers were removed about 20 minutes on the bench saw both oscillators dialed in - and the SSB signal is now "centered" in the filter passband. This means I need to go back and check at least one of the other rigs. :roll:

[N8YX]

This means I need to go back and check at least one of the other rigs.

The local oscillators in three of four were off frequency - some by a slight amount; the others by as much as a kilohertz or more. Setting them was a simple process, albeit time consuming.

I learned something in the process - a nifty method to incorporate a variable BFO without major modifications to the 3rd LO circuit: Let's just move the 1st LO. Set points for USB and LSB carrier frequency are implemented via a pair of trimmer potentiometers; let's put those controls on the front panel in the form of a ganged pot and provide an adjustment so that "12 o'clock" = centered.

This mod and a variable AGC via front-panel knob are projects for another day. That day will come when I can find a 5 megohm pot of the same style as those used in the IC-751A/R-71A's ganged controls.