Wednesday, February 8, 2012

QRP...easy as 2N123

I recently put together an 80m CW transmitter using a pair of 2N123 germanium transistors, both of them bearing July, 1959 date-codes. The first stage is a Colpitts-type VXO. A 3.58MHz ceramic resonator tunes from 3501 to 3560kHz using a series-connected variable capacitor. A second 2N123 transistor is configured as a Class-C, common-base power amplifier. The RF output is ~50mW, although it falls to 45mW at the low end of the band.

The basic transmitter schematic is shown below; the transmit-receive keying relay does not appear in this drawing.


Part values

C1: 27pF silver mica (s.m.)
C2: 330pF s.m.
C3, C4: 150pF s.m.
C5, C7: 48nF
C6, C8: 430pF s.m.
CV1: 15 to 230pF air-variable
R1: 100k
R2: 1.5k
T1, T2: 3.5 to 6uH slug-tuned; 16-turn primary, 3-turn secondary
X1: 3.58MHz ceramic resonator
Q1, Q2: 2N123 germanium PNP transistors

Adjusting T2 for a peak RF output, the highest harmonic energy is -35dBc (the 2nd harmonic). However, by tuning slightly off-peak, the peak harmonics drop below -45dBc. Thus, no further filtering is required. 

I've paired this transmitter with my Cat's Meow galena and cat's whisker receiver. While receiving, the RF output signal from the continuously running transmitter is relay-switched into a 50 Ohm cermet potentiometer working as a dummy load. A sample of this signal is taken from pot's wiper to provide the receiver beat-frequency oscillator (BFO). A single, miniature DPDT relay is keyed in order to provide the required switching. A reed-relay was initially used to switch a 10pF capacitor across C4, for the purpose of providing a receive-increment-tuning (RIT) frequency-shift. However, I discovered that with careful tuning the variation in transmitter loading is alone sufficient to provide a RIT frequency shift of 750 to 1200Hz (from the bottom to the top of the band).

At G3XBM's prompting I measured the minimum detected signal (MDS) level required at the receiver. Using a factory-made germanium diode in place of the galena and cat's whisker, I could just copy a -118dBW signal. This is an absolute power of 1.5pW, or 8.75uVrms across a 50 Ohm resistance. Holding the input step-attentuator at this setting, I re-installed the galena and adjusted the cat's whisker until I could just hear the signal in my headphones. For the same signal copy that I had with the germanium diode it was necessary to switch a 3dB pad out of the circuit. Thus, the best sensitivity that I could wring out of the galena was -115dBW. This is an absolute power of 3pW or 12.4uVrms across a 50 Ohm load. 

It's interesting to review a similar measurement that I made for my 80m Reggie transceiver (the receiver was also gain-less and I used the same pair of headphones) back in 2009. I then measured an MDS of -120dBW; which is an absolute power of 1pW, or 7uVrms across 50 Ohms. Allowing for the subjectivity involved in making these measurements, the results of both tests when using commercial diodes are very close indeed.
 
Here's a photo of the "finished" 80m station. The transmitter appears on the rear proto-board, while the "crystal receiver" is up-front. Again, the small green and black reed relay appearing on the left-hand side of the transmitter has since been removed ("Simplify, simplify..." Thoreau :-).  



In two evenings I've made eleven QSOs with this setup.

KA1ANW   579/599 (him/me)   Sharon, MA          225km
WA1AR     579/599             Wrentham, MA          255km
N2UU        579/449               Voorhees, NJ          504km
W1FL        559/339                    Stowe, VT            46km
W3NW      569/539           Lock Haven, PA          504km
KT8R         559/229                        Alto, MI        1021km
W3ZT       579/559            Skaneateles, NY          319km
W1PID      579/589           Sanbornton, NH          112km
WA3MIX   579/579           Williamsport, PA          469km
K1ESE       589/559             Waterford, ME          163km
N8ER        569/559                Parsons, WV          797km

My QSO with W1PID lasted nearly an hour. Jim began with an ICOM 7000 at 2watts, but he quickly switched to his Elecraft K2 in order to play the "QRP Limbo" game. Jim reduced his power in steps: 500mW, 200mW and finally 100mW. His 100mW signal came through my cat's whisker at a perfectly readable 529! I received a long transmission very nicely at 200mW, aside from a bit of QRM towards the end.

I particularly enjoyed my contact with W3ZT on account of his unusual transmitter. Joel was running a 1950's-vintage URT11, which was built for the CIA and commonly used to provide CW communications for our foreign embassies. We had a solid contact lasting 15 minutes.

I admit that I get a kick out of the responses when I explain my setup. For example:

N2UU sent, "TELL ME YOU ARE JOKING."

WA3MIX said, "MAN TTS ABT AS OLD SCHOOL AS U CAN GET = SOUNDS GREAT WID GUD KEYING ES VY STABLE."

K1ESE replied, "DONT THINK I EVER WRKED A GUY ON AN XTAL SET FER RX HI = PRETTY AMAZING"

I've heard a number of DX stations as well in the last two nights. ON6AB, OZ3FD and DL2WR had particularly strong signals here.

2 comments:

  1. This comment has been removed by the author.

    ReplyDelete
  2. The RX sensitivity is very impressive Michael. I must see what can be achieved with my hi-Z crystal earpiece. Best I achieved was -55dBm (-85dBW) as an envelope detector on AM although with the diode being hard switched by a VFO I suspect quite a bit better should be possible.

    ReplyDelete