Final Report on the NorCal 20

The NorCal 20 has some impressive features, including a clean, reasonably stable VFO, built in keyer and audio frequency "read out," good selectivity, and very good second order IMD dynamic range. And we have nothing but admiration for the hard work of all the volunteers who made this complex project happen.

But we were troubled by the outcome of some of our most important tests. We tried to go the extra mile, by getting a second sample, and running the key tests multiple times. Unfortunately the tests kept suggesting the same thing. Somehow, the trade offs that were designed into this radio were not working.

By going to a diode ring double balanced mixer, the designers willingly took the penalty of high current consumption in order to achieve a "crunch proof" front end, with excellent third order IMD dynamic range. But our tests suggest that this goal was not achieved. On top of this problem, our samples were a little hard of hearing. We keep preaching not to make a horse race out of sensitivity, but our samples were among the least sensitive radios we've encountered.

After we published our review, we got in touch with Dave Fifield, the radio's designer, and Doug Hendricks, representing the NorCal QRP Club, the radio's sponsor. We sent one of our samples to Dave, who was kind enough to give our sample some fine tuning. We are delighted to have the opportunity to publish the "mods" appearing below.

Dave send our sample back, and we tested it again. Here are the key results:

• Sensitivity improved from an MDS of -129 to an MDS of -135
• Blocking dynamic range improved from 118 dB to 123 dB
• The third order dynamic range improved from 84 dB to 86 dB
• The third order intercept point dropped from -3 dBm to -6 dBm

We have revised our comparative charts for these four tests, and in the future they will reflect the foregoing results for the NorCal 20. The charts already published between January and April reflect our original results.

******

Here are the comments we received from Dave Fifield:

1. Neither of the two tuned circuits in the RX front-end filter were tuned optimally. They were both set with the trimcaps at or very near maximum value. Instead of messing with the inductors, I simply changed C16 and C18 to 56pF (was 39pF) and retuned the front-end. I now get two nice clean sharp peaks in the tuning of each section in the mid-range of the trimcap values. It made a lot of difference to the RX immediately!

2. I also noticed that L6 and L7 still had the "old" number of turns on them. I took two turns off each (so that L6 = 8t and L7 = 7t). This further improved the RX performance (less loss in the harmonic filter!) as well as making the TX output a bit cleaner.

3. I also noticed that the RX audio bandpass filter (post-product detector) still had the "old" values in it, setting the bandpass peak to about 760Hz. I changed three resistors, R53 to 3.9K, R54 to 7.5K and R55 to 270K. This is a mod that has been around for some time now too. This brought the audio response peak down to about 650Hz which results in a much nicer to listen to RX and also gives it a bit more gain.

4. I also experimented with the match between the front-end filter and the J310 RF pre-amp. I found, much to my delight, that 4t on the secondary was NOT optimal, nor was 3t as Dave Meacham has postulated, but in fact 6t on the secondary of T2 seems to be optimal.

The net result of these few simple changes gives a MUCH nicer performance. As I mentioned, I measured the MDS as -133dBm now, and comparing it to my TS-950SDX, it hears everything that the big rig hears, no problem.