Experiments with the Sony SRF-59

I got my start in the radio hobby listening to a Japanese 8-transistor radio that my aunt bought me for Christmas in 1962 or so. Up to that time I don't think that I had ever even tuned a radio. We had an AC/DC metal-cased deathtrap AA5 in our kitchen, but I was discouraged from messing with it. In those days you were liable to hear some pretty odd things on AM radio at night, and of course the magical things that happened to make those distant stations heard made a lasting impression.

Even as I moved on to ham radio I kept an interest in listening for DX on the AM band, and I always had a little portable radio that I would often listen to in bed before going to sleep. Some years back, I had a Sony Walkman-type AM/FM radio, and it died, so I replaced it with a new Sony SRF-59. The radio worked OK on FM, but it was pretty deaf on AM. Quite a disappointment, but I lived with it for a few years until the tuning capacitor started making noise. I threw it in the junk pile and looked around for a new replacement, but Sony was still selling the SRF-59. I tried some other cheap radios but they were all pretty bad, with lots of images (I hate images) and poor sensitivity and selectivity. Then I read on the Internet that the SRF-59 was considered by some to be a hot radio - really hot, in fact.

I took a chance and bought another one (actually, two) and I was sort of surprised to find that it really was a hot radio and the new example was nothing like the old one I had tossed. Some reviewers on the Internet said that the SRF-59 had absolutely no images, but I found that there were noticeable image responses to at least two nearby transmitters. Other than that, the radio was pretty much as advertised, with very good audio, selectivity, sensitivity and AGC action.

Looks cheesy, but it's a good AM receiver

This all got me curious about the radio I had thrown into the pit of despair. Its variable capacitor gang was definitely TU, but maybe the rest of it was OK and worth screwing around with. I retrieved it from the pile and tore it apart. It looked to be the same board referred to in the SRF-PSY03 service manual that is available on the net. Snooping around the schematic gave me a couple of ideas. I removed the variable cap and the on-board padding caps C5 and C15. I wired the oscillator section of an old AM radio tuning gang, one with a nice reduction drive, to the AM OSC pad on the '59 circuit board. I wired up another 365pF or so variable cap to the AM ANT pad. So now I had two separate tuning knobs, but if this lash-up worked that wouldn't bother me too much. I'm pretty used to homebrew and commercial receivers with separate preselector tuning controls. One advantage of this setup is that any "alignment" occurs when you tune in the station, so it is always perfect. Since the capacitors don't track, you just can't have any tracking problems!

If you look closely, you can see the trimpot on the PC board

Anyway, I powered the monster up and it worked just fine! It felt like a different radio because of the dial reduction drive, but the oscillator cap tuned the radio from about 500 to 1800kHz, so that was good. The preselector cap covered the range, though 365pF turns out to be too much capacitance and it's only about half meshed at 530kHz. The antenna circuit tunes very sharply, and on the low end of the band you can hear the selectivity advantage when you have it tuned dead on. I wonder if the tracking on a stock radio can ever be this good. This old radio was definitely at least as sensitive as the newer example of the '59 that I had, and otherwise apparently identical in performance.

Even the images were identical. I noticed on the schematic that pin 26 of the CXA1129N is labeled "IMAGE ADJ." It is connected to B+ through R2, a 240 ohm fixed resistor. Hmmm. This chip implements a superhet radio with an image-reject mixer and active filtering at the 55kHz IF. Image rejection in this scheme is dependent on the balance between two mixers in the front end, and I was guessing that this resistor is what trims that balance. The image responses I was hearing were 110kHz below the frequency of the station causing the image, and there were two pretty good ones, as I said. I replaced R2 with a 1 kohm trimpot and tuned the radio to a station that was experiencing whistle from one of the images. By adjusting the trimpot, I was able to reduce the image whistle by at least 6dB, probably more. The end result was that one of the two original image carrier whistles was now below the noise level and the other one was still audible, but not objectionable, even to someone who hates images.

So what I ended up with was a pretty good receiver for the AM band. The mechanical configuration of the creation was dictated by what was easy to do with hot-melt glue in a few minutes, so it's not too convenient and the metal in the variable capacitor is probably interfering with the loopstick antenna. It's almost as if the pattern is now unidirectional. The image rejection is superior to that of the newer radio that I have. I'm wondering if the mixer balance is dependent on battery voltage such that image rejection will change as the battery ages.

Anthony Felino (AF) WN6Q
PO Box 2702, Santa Barbara, CA 93120
(805) 216 5287