Description

Homebrew DX Contest 2007

Tim Kilboy
Jackson, TN
The Direct Reflex is a unique design which builds on the 1920's technique of reflexing RF and AF through a single grid on a triode.

In the old triode reflex receivers, a 2nd tuned RF tank was placed in the plate circuit to capture amplified RF. This signal was then rectified and passed through an AF transformer. The AF transformer secondary was then put in series with the 1st tank, in the grid circuit. Now both the original RF as well as the detected AF are both impressed on the grid. The set is now a two stage receiver using one tube.

This old design has the distinct advantages of good 'two stage' sensitivity, and good selectivity due to being double tuned, since it incorporates two tuned tanks. On the other hand, the second tank was not very selective since it was feeding the relatively low impedance of the 3:1 AF transformer primary.

Reflexing died out fairly quickly since it was only really effective on single tube circuits. Although it was tried on 3 and 4 tube sets, to get double the gain, there were endless problems with keeping the set stable across the frequency range that were difficult to overcome. So since reflex technology was best suited for one tubers, and the market was quickly moving toward 5 tube neutrodyne sets, reflexing became extinct.

Enter the 'multi-grid' vacuum tube. During the reflex heyday, only low gain triodes were available and by the time the screen grid tube came along, reflexing was a forgotten fad. However, taking a more modern tube like the 12BA6 Pentode RF Amp with its higher gain (4400 uMhos) and using its suppressor grid as a second input grid (for audio), provides a new spin on an old idea. The anticipated benefit is that both tanks may now run with very light loading, which should lead to optimal sensitivity and selectivity.

The 'Direct Reflex' set demonstrated some excellent traits, but also had some problems. The performance can be broken into ranges. 530-1250 KHz - Superb sensitivity and selectivity, except for 850-950 which was somewhat 'dead'. 1400-1700 KHz - reflexing was not as effective, seemed to work best with reflex gain turned very low. Tube selection may be an issue, since the 12BA6 specs the suppressor (grid 3) as also being an internal shield. Use of the suppressor grid is probably not the optimal way to get a second input. The set could be pushed to oscillation at high gain settings. I improved this problem by putting the second tank in the cathode rather than the plate circuit. Even with this, it seemed apparent that the suppressor grid was 'talking' to the No 1 grid - perhaps due to the shielding function of the suppressor.

Despite its shortcomings, the Direct Reflex made a very respectable DX showing. Station separation was very crisp and sensitivity was as good enough to make it impossible to find a dead frequency. Building this set has probably raised more questions than it answered. I think with selection of a more appropriate tube, it would make it a home run.

Pictures