One Tube double-Reflexed Superhet Receiver!
My fellow collecting friend Barry Kassindorf, just sent me a link to an amazing recent home brew radio:
This radio accomplishes the function of a full super heterodine radio with just one 6ME8 sheet beam tube.
Robert Weaver designed this radio and has kindly consented to the post of the schematic here, but his web site site above is well worth a visit.
RF is first ampllified by the control grid of the tube, then after the sheet beam passes the post-acceleration grid, the RF is converted to 455KHz IF.
A completely differential Local oscillator takes output from the differential plates and regenerates it via the differential deffection plates.
The just-converted IF signal is also extracted differentiallly from the differential plate output by a carefully ballanced differential IF transformer. There is capacitive trim dedicated to maximzed common mode signal rejection at the grids.
This Balance IF transformer has an unbalanced secondary output that drives ceramic 455kHz IF filter, which, in turn drives an idential winding as the primary of the next IF conventional single-ended IF transformer.
After all this muilti-pole IF preselection, this IF transformer reflexes the IF signal back to the same control grid that first passed the RF signal input.
Now the IF signal is amplified as a common mode signal, comming out with identical phase and amplitute at the two output plates. This common mode IF signal is harvested at the center tap of the first IF transformer (which is carefully ballanced to reject this common mode IF signal) and tuned by another single ended double tuned IF Transformer.
Finally, the secondary of this transformer drives a conventional grounded cathode detector.
DC bias for AGC and audio are separated at the detector and driven independently to the control grid, for a third re-entry of the signal path into the control grid of the 6ME8.
The control grid of the 6ME8 is driven in series by a coil injecting input RF, a coil with IF, and a rejoined combination of the Audio and AGC signals. Four signals in total - this may brake a record!
Lastly, the audio signal travels the same common mode path of the IF signal, with the plates driving the audio transformer through the center tap of the first IF balanced transformer primary, and the primary of the last IF transformer.
This is a truly amazing feat, and I think it merits the creation of the imaginary "Grundig-Reflex prize" for the most exhaustive use of an active device. This design deserves a place along the legendary Three tube AM-FM Grundig Superhet Reflex radios (model 80 was the first) and the single pentode Grundig FM stereo decoder.
This radio and it's creator, Robert Weaver, deserve to be the first glorious recipients of this imaginary prize!
In the late '20ies and '30ies reflex amplification was often used in German receivers due to a fee which had to be payed to Telefunken for each tube base in a receiver. This is also the reason for the development of multi-tubes like 3NF (Loewe).
The above mentioned Reflex Receiver was produced with identical schematic, however different housings by AEG (Geatron 34WLK), Telefunken (Meistersuper 332WLK), and Siemens (37WLK). The receivers were also produced as pure DC sets (not AC/DC).
For the reflex stage a Hexode type tube (RENS1234 resp. RENS1824) was used, which was produced only during a very short period of time. The speciality here is the detection within the reflex stage.
A description of the reflex principle used with these receivers can be found here (in German).
I was very impressed with the diode plate detection at the RENS1824 hexode. This is one very busy tube, and the hexode was a European specialty.
I have in my collecition a couple examples of reflex designs in American radios. One is the FADA160 from the 1920's with 2 RF stages, a detector stage and the first audio stage reflexed into the first RF stage. One last tube drives the horn. all tubes are 01A's. This was also a neutrodyne set.
The other set is a GE 212 AM-FM radio from 1948 with the first IF stage reflexed into the first RF stage. A very conventional reflex implementation.
Speaking of Reflex and counting stages, what is the German method for counting circuits "kreis" in a radio? Do the circuits account for the number of complex conjugate pole pairs in the signal path, or resonant tank circuits involved in selectivity and image rejection? Is the local oscillator included in the circuit count?
One fine point about the alignment of the first IF transformer in the One-tube Superhet at the start of the thread:
The plate balance trim caps are adjusted for maximum rejectioin of the common mode IF signal that is injected at the control grid. One way to state the alignment requirement is that the Common mode rejection must be deeper than the IF (common mode) gain from the control grid. If the Common mode rejection is not deep enough, IF oscillations will result.
The concept of amplifying independent signals in a common mode path and in a differential path was used as a way to implement stereo amplification in a basically mono amplfier structure. I attached an example of this principle, called Stereo Simplex by CBS.
- CBS_stereo_Simplex 1 of 3 (113 KB)
- CBS_stereo_Simplex 2 of 3 (85 KB)
- CBS_stereo_Simplex 3 of 3 (95 KB)
In a private email about this thread, Radio Age Magazine editor Ed Lyon shared his past professional experience in the 1950's with sheet beam tubes being used to amplify signals in common mode and differential mode.
Quoting Ed's email:
At 09:02 AM 4/6/2009, Ed Lyon wrote: Joe:
We have tried separate-signal amplification using a nearly-similar scheme, but with the 6LE8 tube, which is probably easier to find. It has one cathode, one control grid, one screen grid, but two suppressor grids and two plates, I think on each side of the cathode. The Gm for the grid-no. 1 is 5800 uS and for the suppressor grids it is 350 uS for each respective plate.
Otherwise, it should work in a similar reflexed receiver circuit to the one you have shown. Your friend's circuit is a clever arrangement for separate-signal amplification, as that was what we were trying to do (I think it was in the late 1960s)in an ice-floe reporting system. The 6LE8 has a bit more Gm than the 6ME8, as well, and the suppressors work around ground potential, rather than at 75 volts or so, and draw practically no current. The low Gm for these suppressors is of no consequence, either, since one could simply cross-connect the plates and suppressors with little capacitors, for the L.O. feedback path. I would use a three-section variable cap for tuning, with two sections used for the balanced L.O. circuit, and the third section for RF tuning. Lower Gm versions are the 6BU8 and 6HS8.
The ice-floe reporting scheme was needed in the far north in the springtime, when loosening ice floes would wander in the Arctic ocean, and might threaten oil well drilling rigs. With sufficient warning that such a floe was coming close to a drill setup, the drill-pipe and bits could be retrieved and the drill ship could scramble out of the way of the floe, to return later and resume drilling. The reporting system tuned in the VLF signals from the OMEGA navigation system (today we would use GPS) and translated them to HF where they were reradiated when interrogated by an HF over-the-horizon radar. At the radar, the HF signals were recovered, retranslated downward, and detected by an OMEGA receiver to tell us where the ice floe was located. The unit on the ice floe was air-dropped and self-erected an antenna for both the VLF signal and the HF repeater. We later redesigned the whole thing in solid state.
There was another tube very much like the 6ME8 called the 6JH8, which you could look for at hamfests.
I would think you could do a similar multiple-reflex circuit with one of the twin-pentodes like the 6HU8, also called in Europe, ELL80. This is a powerful tube that would probably make more speaker volume. It has a single large cathode and a control grid, screen grid and plate on either side of the cathode. The control grids could be connected together and the L.O. could simply cross-connect the screens and respective plates, tuned by a balanced tank circuit, since both screens and plates are rated at the same B+ voltage.