I've recently converted a Selena B-211 (Океан 209) to the 88-108MHz FM and so I thought I'd write up the process, as several of these radios are out there and the process is fairly straightforward as long as everything is figured out (which I did and hence sharing my experience). Also, other Eastern-bloc radios (Sokol-308, etc..) can be converted in a similar way.

The radio had an export version Selena B-212 with its FM band in the 88-108Mhz range. I am fortunate enough to have this export version available for reference; also, the complete schematic diagrams of both versions are available on the internet. (attached below)

The construction of both (eastern and western) versions of the tuner is fairly similar with only minor differences; these differences were insignificant enough for me to take them into consideration (different placement of the D3 clipper diode on the IF transformer, an extra 12 ohm resistor (R12) in the B-211 version, different biasing technique for the AFC varicap, different biasing resistor values of T2, etc..), so I mostly focused on the tuning- and RF component values.

After visually inspecting both tuners (B-211 and B-212), I concluded that all the main tuning components are identical, except L3, which is the coil of the continuously tuned RF LC tank. In the B-211 (66-74MHz) version, this coil has 6 turns, with a tap at the 4rd turn (counted from the bottom). This coil only has 5 turns in the B-212 (88-108MHz) version, with a tap at the 3rd turn from the bottom. So essentially I had to unwind one turn from the bottom of the coil of the B-211 to make it identical to the B-212.

After making L3 coil shorter (and identical to L3 of the B-212), I basically replaced all the RF capacitors to match the values of the B-212 tuner.

C1 30p -> 10p
C2 68p -> 30p
C4 22p -> 10p
C6 75p -> 150p
C9 36p -> 18p
C13 15p -> 4.7p
C17 18p -> 4.7p
C19 13p -> 6p

I used oscilloscope (at the oscillator coil tap, through a 1p capacitor to the probe) to set the tuning range of the oscillator between approximately 98-119MHz (10.7MHz above). After all this was done, I re-installed the tuner and did a final alignment on it.
The alignment is pretty straightforward: find a mid-weak station at midband (around 98MHz) and peak up the reception with adjusting L1 and L3, using the tuning indicator of the radio. None of the IF tanks need alignment as long as the IF tuning is in factory condition.

Usually this is the point where similar FM conversion works are considered finished, with disappointing results. The radio is harsh-sounding and FM audio is distorted. The poor quality -especially the distortion- is usually attributed to the assumed inaccurate alignment of the radio and the tuner.
However the poor sound quality has surprisingly nothing to do with the tuner, or with the alignment, if otherwise the IF section was untouched. The harsh sound is the result of different pre-emphasis used in Europe and in the U.S. (75us vs. 50us); the distortion is caused by the deeper modulation used on the 88-108MHz band, which overdrives the IF stages and the FM ratio detector of the Eastern-bloc radios.

Bridging the de-emphasis capacitor (C142, 6.8n) across with 3.3n (or replacing it with 10n cap) fixes the de-emphasis and the radio doesn't sound harsh any more.

The overdrive comes from the difference in modulation level of the two bands. The Eastern-bloc 66-74MHz band used approximately +/-50KHz modulation depth and hence the narrower-band IF stages and the sharper FM detector of these radios distort with the +/-75KHz modulation depth, which is the standard on the 88-108MHz band. Since re-designing and re-building the entire FM IF strip is impractical (yes, the B-211 and B-212 radios have different IF strip and detectors, due to the different modulation depths), a relatively easy and effective solution is modulation compression, i.e. a small portion of the audio signal is fed back to the AFC varicap diode, which then works against the FM modulation and hence reduces the frequency deviation that is reaching the IF stages and the FM detector.

I added an audio signal path in parallel with the existing AFC path. It's important that the audio is taken directly from the detector, i.e. before the de-emphasis network and any filtering. The two added 47k resistors form a voltage divider between the ratio detector and the ground (whether the AFC is switched on or off, via C143 and the added 100n cap). The added 47p capacitor keeps RF out of the loop. The 47k resistor on the detector side is soldered directly to one of the leads of the electrolyte buffer cap (C141) of the ratio detector, in order to get 100% pure detector output.
Sketch attached. The top drawing is the original AFC network of the radio, the bottom drawing depicts the extra audio feedback for the modulation compression, with the added extra components circled and marked.
The added resistor values should be as high as possible in order to minimize the load on the ratio detector, but also low enough so that the low-pass filter formed by the added resistor (in this case 47k), R10 (56k) and the varicap itself (somewhere in the 10p range in this case) still allows the full audio bandwidth to modulate the varicap.

The overall result is decent FM reception and enjoyable audio quality on the 88-108 band.

Hi Szoftveres,

Very interesting insight, thanks for sharing it!

I also have a B211 radio and also in the process of converting the FM to the CCIR band.
From the "factory conversion" it seems that while the OIRT version has high-side injection, the CCIR version has low-side injection.
That is also the possible reason for the differing IF-boards. If you notice the ratio detector diodes and that small electrolytic cap in the detector are reversed for providing the "negative feedback" for the AFC circuit (assuming the coils in the detector circuit have the same phasing).
May be you could confirm this with the your factory CCIR version radio?

The first attempt with this factory conversion did not work for me, the frequency spread just was not there.
I have given up on that mode, I also wanted to have high-side injection, I was not looking forward to modifying the IF-board too.
I also rewound the osc coil, removed turns from it (not sure how much, that conversion is an ongoing on/off effort).
That is when the conversion process went into a series of laughable blunders (Murphy is always with me), but finally, I (hope I) know what is going on.
The last roadblock that I am facing now is that the variable cap has fallen apart, first there were some intermittent shorts in the osc section, and now the stationary plates have let loose, the crimping to that ceramic post at the bottom got loose.

Very interesting comment on the effect of the higher CCIR deviation, and resolving its effect!
I have not used the rest of the radio yet, I have an old FM radio kit that I only built partially. It uses a TBA120S FM-IF chip and some obscure TI audio amp chip. As it only had a single FM-IF filter can, I added a ceramic filter with a JFET amp stage.

There is a very entertaining "essay" about this radio here (other items in the series are also interesting/entertaining):
http://szetszedtem.hu/065selena/taskaradio.htm

PS: The tuner schematic you gave is for the Ocean-209, the B-211 is slightly different (at least in mine), the AFC diode has separate bias resistors.

Regards, Peter

My B-211 (the subject of the original post) originally used high-side injection (I measured the oscillator frequency somewhere between 75-86MHz on the oscillator coil tap). I can easily imagine that these tuners evolved over time and had different revisions - with both high- and low side injection (my high-side injected B-211 is from '79).

The ratio detector diodes and the AFC varicap are indeed reversed in the B-211 and B-212, however this has nothing to do with the low- and high side injection (both my B-211 and B-212 are originally high side injected), it's presumably just a consequence of some unrelated design decision which I wanted, but haven't figured out yet.

I verified both my B-211 and B-212 tuners against the schematic by checking the capacitor values printed on the caps, then made a list of what needed to be modified in order to make the B-211 tuner RF equivalent to the B-212. The most significant change was the unwinding of one turn from L3 (far right, just in front of the tuning cap), the rest was just capacitor work (almost all the RF cap values need to be lowered, with the exception of C6, which needs 150pf in place of 75pf). My B-211 and B-212 use the exact same oscillator coil, the capacitor values make the difference.

The sharper nature (and therefore easily overloading) of the IF and ratio detector in the B-211 are apparent from the different capacitor values used in the IF and ratio detector LC tanks used in the B-211 and B-212. Modulation compression is one way of solving the problem when the AFC diode is available; lowering the Q of each LC tank in the IF section with a shunt resistor is also a solution (more suitable for tube radios), however that would necessitate finding the right resistor values, ugly soldering at the bottom of the PCB, and then re-aligning the entire IF section - these radios retain their factory IF alignment quite well, so I didn't want to go in that direction.