Sunday, October 21, 2012

Vacuum Fluorescent QRPp

I received a Tung-Sol 6977 vacuum fluorescent indicator last week from a friend. Wow...this is the smallest vacuum tube envelope that I've yet seen! The above image of an illuminated 6977 was made by Joe Sousa in this informative post. A quick search for early documentation turned up this 1958 article (pp. 5 to 9 ). Naturally, I wondered whether it might be possible to build a radio from this unusual device. 

I used my 6977 to build an 80m regenerative receiver the next day. At first the circuit refused to oscillate, despite having a rather high level of feedback. Temporarily opening the feed-back loop, I noted that it was indeed acting as an RF amplifier. Closing the (positive) feed-back loop again increased the gain, as expected. Obviously, the loop-gain was below unity. Increasing the grid-leak resistor value from 3.9 to 10MegOhms did the trick; adjusting the regeneration control (see schematic) now allowed the receiver to slide smoothly in and out of oscillation. The receiver tunes from 3500 to 3583kHz. I spent the remainder of the evening logging DX stations at the low-end of the 80m CW band. 

The next morning I experimented with an 80m, crystal-controlled, Miller oscillator using my 6977. Working to maximize the RF output, I eventually measured slightly more than 1mW across a 50 Ohm load resistor. Given the 57Vdc anode supply at 320uA, this represents a dismal (~6%) RF output/DC input power efficiency. 

The anode to antenna impedance matching network is to blame. If the efficiency were a more typical value (say, 25 to 30%), I should expect to measure 5mW of RF output power. In this case, the standard rule-of-thumb formula indicates the anode should be presented with a load impedance of (57^2)/(2*0.005) = 325kOhms.

The most suitable output tank inductor in my junkbox was a "Hescho"-type (14.7uH, Qu ~250) ceramic/silver unit that I picked up at the hamfest in Friedricheshafen. The unloaded equivalent parallel resistance of the resulting output tank resonator is (250*2*Pi*3.5^6*14.7^-6) = 80.8kOhms. This is too low to efficiently produce the required impedance transformation. 

The physics of this matching network also implies the tuning will be very sharp. On the other hand, harmonic energy is attenuated quite handily. In fact the second (and worst) harmonic is down by 43dBc, so at least no further spectral filtering is required.

It would be interesting to try a crystal-controlled Colpitts oscillator configuration (raising the filaments up from RF ground via an RFC). Taking the RF output at the filaments ought to ease the required impedance step-down ratio somewhat.  

Vacuum Fluorescent QRPp
Component List

X1: 80m quartz crystal
V:1: 6977/DM160 vacuum fluorescent indicator triode
R1: 100k
R2: 10MegOhms
C1: 20nF
C2: 2.5nF s.m.
C3: 5pF s.m.
C4: 360pF s.m.
C5: 40pF s.m.
C6: 100pF s.m.
C7: 3.9nF
C8: 100uF @ 100V
CV1: 20-335pF air-variable
CV2: 15-420pF air-variable
CV3: 30-360pF air-variable
L1: 14.7uH, high Qu (>250)
L2: 3.3mH
T1: 3.5 to 6uH, slug-tuned "IF can"
T2: audio transformer, 13.2kOhms to 600 Ohms
RY1: DPDT "DIP" relay, 12V coil
D1: 1N4148, 1N4002, etc. 

I decided to build a transmitter/receiver by relay-switching the grid and the anode of a single 6977 triode between the receiver and transmitter circuits. This worked out quite well. As I often do with my regenerative receiver/quartz-crystal transmitter pairs, transmit frequency spotting is accomplished by listening for a tiny "blip" or "ping" in the headphones as one tunes through the series-resonant frequency of the crystal. 

When everything was ready to go, I ran up to the house to post a note on the QRP-L mail reflector with the details of my intended operation. Once on the air it became apparent that it wasn't a good night for milliwatt work; the high QRN was simply masking my signal. 

Nevertheless, I called CQ for several hours. Sometime after I'd resigned myself to the fact that I wouldn't make a contact, there was W1PID calling me! Jim remarked on the QRN, but he also handed me a 449 report. He was a perfect copy on the regenerative set; I called it a 579. What fun! Working a station 112km distant with 1mW through heavy 80m QRN augers for "big DX" come the quiet winter months. 

I made a short video of my Vacuum Fluorescent rig shortly after my QSO with W1PID.

Link: MØAYF, builds regenerative receivers from surplus VFD's. Well done, Des!

Mike, AA1TJ                 



  1. Nice work Mike!

    I must try a project with valves (tubes!)myself soon.


  2. Hello Michael, Impressive design with a small number of parts. I like your video. Great to see that your video also has a part with the lights out. hi. The keying is very fast. It's great fun to use such a divice. Thanks for this very special Blog entry. 73, Bert

  3. Hi Bert, Thank you for your very nice comments!

    Mike, AA1TJ

  4. I have a pair of these in my collection and was researching the exact same idea when I stumbled on your blog! Very exciting, great work! Thanks, I'll get this on 80m asap and see if I can replicate your success. KC9TUI

  5. Hi John, Aren't those little 6977's cute?! Good luck with the project and please let me know how you get on.


  6. I have written a few web-pages recently. You might like this one:

  7. can you do this with nixie tubes? they are plentiful but a quick web search has offered no results.

  8. Thank you for the interesting info on JFET regenerative receiver circuits.

    Kind regards,

  9. Dear Stephen, I'm afraid not. Nixie tube operate by ionizing a back-filled neon gas mixture. The ionization/de-ionization times are long in comparison to the cycle time interval (1/f) at HF.

    That said, just for grins I once pushed a neon lamp relaxation oscillator frequency to the upper limit (~8 to 10kHz). The resulting pulses were used to "ping" a high-Q LC resonator at 600kHz. I recall being able to receive the raspy signal around my shack on a portable BCB receiver. No doubt, the spectral purity was akin to an ancient spark transmitter. :-)