Wednesday, December 28, 2011

A Swedish Sputnik and More

Johnny, SM7UCZ has made a top-notch job of his new Sputnik transmitter!

The schematic diagram of his circuit layout, plus more photos may be found here.

Among the many interesting tube and transistor QRP projects to be seen on Johnny's web site is the story of a hand-transcribed copy of Frank C. Jone's, 1938 edition of, The Radio Handbook (please click here)

OM, Sven Bernholm, SM3RN (1888 - 1963), laboriously translated this work into Swedish whilst standing radio watch in the years 1944 to 1945. Sven beautifully hand-copied over 400 pages of text, schematics, drawings, graphs and tables. I'm surprised that he didn't illuminate the pages in the fashion of medieval monks! His little side-notes are especially interesting; for example

5/5/1945: "Denmark and the Netherlands, free!"

7/5/1945: "Germany surrendered; Norway free!"
12/4/1945 "Roosvelt dead - Allies 6 miles from Berlin!"

Johnny recreated a Frank Jones-type "push-pull" transmitter similar to the one appearing in Sven's transcription (here).

There are several Parasets, a Hartley transmitter, a tube superregenerative receiver for the VHF BCB and a transmitter built to operate on 1.5Vdc (and less). The fine art of homebrew QRP is alive and well in this little corner of Sweden!


Wednesday, December 21, 2011

The Cat's Meow

"Who would not, after once trying it, admit that the three element vacuum detector is the most sensitive device known for the reception of Herztian waves? The author freely grants this, but claims that for pure, unalloyed pleasure one must experiment with the less sensitive detecting devices before one can know the keen delight of the radio game." Lloyd Manuel, QST Magazine; July 1916, "What's the Matter with Galena?"

I got the "galena" itch again yesterday. Luckily, it's an easy itch to scratch as no more than five minutes are needed to lay down a heterodyne receiver on the protoboard. My circuit is shown on the left in the drawing below. For comparison, the heterodyne receiver used for long-distance radio testing aboard the USS Salem in 1913 appears on the right-hand side (please click-on the photos in order to enlarge them).


Component values for "The Cat's Meow" on 80m

C1: 510pF
C2: 2.8nF
T1: 3.5 to 6uH, slug-tuned, 16-turn primary, 6-turn secondary
HP: 600 Ohm magnetic headphones
BFO: I used my bench RF signal generator set for ~ 1Vpp

Not shown in my schematic diagram is a 10nF capacitor connected between the mixer diode and the BFO. I simply forgot to include it in my drawing.  

The mixer diode is a galena and cat's whisker; a gift from my pal, DL3PB. The receiver appears to the right of the R/T toggle switch in the photo below.

In order to get on the air as quickly as possible yesterday I rebuilt my old Vermont Snowflake transmitter. The VXO'd 3.58MHz ceramic resonator allows me to cover most of the 80m CW band. I used slightly different component values for the transmitter low-pass filter yesterday, but it was the same circuit otherwise. The RF output power is 17mW into my 80m half-wave wire, end-fed wire antenna.

It would have been possible to use the transmitter output for my receiver BFO, but that would have meant including a provision for receive increment tuning (RIT). I decided to keep everything as simple as possible yesterday. The manually operated SPDT switch shown in the photo below transfers the antenna between the transmitter and the receiver. A DPDT switch would have been better as I could have used the second pole to switch-off the transmitter oscillator while I am receiving. The complete setup appears in the following two photographs.

I made my first QSO literally minutes after becoming QRV. I answered WA1HFF's CQ and he came right back to me. I made a total of three QSOs for the evening.

WA1HFF  579/449  Longmeadow, MA  141mi/227km   50w/dipole
K3SEW   579/439  Howard, PA           326mi/524km  
N1MX     579/439  Walpole, MA         154mi/247km   100w/dipole

When I came up to the house I found an email message from John, KQ1P

I just listened to your CQ QRPP. I copied you 439 here in ME. I called you at 1 watt to my 80M loop = NIL. Went for 7dB increase at 5 watts = NIL. Your signal sounded good, some QSB. Time to retire now, will listen for you tomorrow if you try again. What new experimental rig are you using?

Gosh, I should have heard him, even with 1 watt! We're set to try again tonight. 

Here is the list of Reverse Beacon Network captures of my signal last night.

22 December 2011: 

KQ1P  219/449  Orland, ME  207mi/331km  5w 

Today's RBN log of my 17mW signal

My affection for the galena and cat's whisker radio began when I was a Cub Scout in the mid 1960's. I remember being intimidated by the sight of this circuit in my Wolf Scout handbook. One of these days I really ought to build this radio. A few minor changes and the addition of a BFO for CW reception would make it into a 1st-Class DX machine! 


Thursday, December 15, 2011

W1PID's Christmas Tree DX Adventure

December 23, 2002

It was the day before the night before Christmas and all through the
house not much was stirring. The girls had gone out for some last
minute shopping.

I had this goofy idea...why not load up the lights on the Christmas
tree in the living room for an antenna? Nice holiday QRP experiment...
So I gave it a try. I worked Norway on 20 meters and Scotland on 15
meters with 5 watts and a little tuner.

When I told Fred, GM3JKS about the setup, he sent back:
"Congratulations on the 5 watts and Christmas tree lights for antenna. This is a first. Hope your family is impressed." I told him they were out shopping and they probably wouldn't be a bit impressed!

The whole thing was magic, but not on the scale of Santa, his sleigh
and the coming down chimneys. It was, after all, thirty feet or so of
wire wound spirally around the tree. 100 lights in all. And they didn't
glow or anything. I'd hooked the tuner up to the ground side of the
plug, and I'd strung out a 20 feet or so of counterpoise along the
floor. I'd worked England from the dining room table before with a 3
foot antenna, so I knew it was all possible.

But it was fun. More than was a little Christmas ham magic,
and I think I'll try it again next year.

Happy New Year from Jim W1PID.

Thanks for allowing me to post this, Jim! 

A Saturable Core Polyakov Mixer on 80m

Vladimir Polyakov (Вертикал Полякова), RA3AAE (on left)

This 80m "Polyakov" direct-conversion "proof-of-concept" receiver uses a pair of vintage ferrite computer memory toroid cores in the subharmonic mixer.  A 1.75MHz beat frequency oscillator (BF0) drives the ferrite cores along the full "B/H curve." The cores are driven alternately in and out of saturation twice per working cycle. 

While the cores are saturated the inductive reactance falls to a low value and the signal from the antenna passes through the mixer with little attenuation. While the magnetic flux is passing from one saturation state to the opposite state the inductance rises (ideally) to a relatively high value. This high reactance momentarily blocks the antenna signal's path through the mixer. The idea is to create a commutating mixer from a time-varying, series reactance. 

I'd hoped that it would be possible to hear 80m CW signals with only a pair of headphones connected to the mixer output. Unfortunately, the conversion loss is presently some 35 to 40dB higher than is typically seen with a diode mixer. 

The mixer ON/OFF ratio ought to be roughly unity over the course of an incoming signal cycle, whereas my saturable reactor mixer spends spends far too long in the ON (saturated reactance) state. Another issue is the maximum reactance obtained in my present setup is too low in relation to the 50 Ohm path impedance. 

In order to measure the "ON" feed-through signal loss an experiment was performed with the BFO signal replaced by a DC source. The current was raised until the inductors saturated. In this experiment I used a variable capacitance in series with the mixer signal windings in order to better cancel the residual inductance. I measured 1.46dB of signal attenuation. No improvement was seen with an RF choke placed in the BFO windings. This tells me the twin core signal cancellation technique (due to reversed polarity windings) is working well.

It wasn't quite so easy to measure the unsaturated feed-through signal attenuation under time-invariant conditions. When the DC bias test current is switched off the saturated ferrite cores simply revert to their remnant state. This is exactly what you'd expect of a semi-hard ferrite computer memory core. It's only possible to get an idea of the maximum reactance by reversing the DC bias polarity and slowly increasing the current; watching the 3.5MHz signal for the point of maximum attenuation. I measured a maximum total attenuation of 9.34dB. This "best-case" differential attenuation is only 7.88dB; which helps to explain why the conversion ratio is so poor.

This mixer working alone into the headphones (no semiconductors in the signal path) produces an audio beat-note with a few hundred millivolts of 3.5MHz signal injected at the antenna input terminal. Again, this is not exactly stellar sensitivity.  

So much for the bad news. The good news is that I was able to copy a fair number of stronger 80m on-air CW signals using this mixer plus one stage of audio amplification. I used a 2N35 NPN germanium-alloy transistor that was made by Sylvania in 1956. Early evening code practice sessions from W1AW are easily copied anywhere in my (admittedly, tiny) shack with the headphones resting on the workbench.

I was careful to insure the input to the transistor amplifier is free of RF energy so there's no doubt the saturable reactors are providing the RF mixing action. Please notice the lack of a band limiting filter between the antenna and the mixer input. I expected this mixer would behave well in the presence of strong adjacent and out of band signals, and so it does. 

My remote antenna coupler provides a return path to ground for audio frequencies. Lacking this, it would be necessary to connect a radio-frequency choke across the receiver antenna input terminals; something in the range of 1 to 2.5mH should work well.    

C1, C2: 1uF
C3: 100uF
C4: 330pF
R1: 330k Ohms
L1: ~3mH 
T1: 50 Ohms to 1500 Ohm AF step-up transformer
T2: 10k Ohms to 600 Ohms AF step-down transformer
HP: 600 Ohm magnetic headphones
Q1: 2N35 (a mid-1950's vintage germanium alloy transistor)
SR1, SR2: 50 turns on signal winding, 10 turns on BFO winding. 

The semi-hard, ferrite cores used in my saturable reactors, SR1 and SR2, each measure 1.27mm diameter by 0.395mm in thickness. They were made in the 1950's-70's for computer magnetic memory use by Kombinat VEB Keramische Werke of Hermsdorf, in the old DDR

By the way, my saturable reactor mixer has much in common with a sensitive flux-gate magnetometer. Sure enough...a small magnet waved within 30cm of the mixer produces a curious "scratching" sound in the headphones. The same magnet set in place within 3cm of the mixer blocks all mixer action. I expect the external magnetic field simply prevents the BFO from effectively driving the core flux around the "B/H loop." 

Here's a photo of the experimental receiver. The ferrite-based mixer appears on the left. The 1.75MHz BFO is in the background. 

Here's the original container filled with tiny magnetic computer memory cores.