My 120-mile Hike Carrying Homebrew Tube Transceiver

Powdered-iron cores (for compactness and relatively high immunity to loss coupled in from surrounding objects, plus immunity to unwanted coupling to other coils) have been around for over a half century.

 

four figures worth of pF in the tank will disprove that "Selectivity" in an oscillating regen is a thing, but only so much so. I will say that the closed-loop rolloff does help kill AF highs -- with too-light coupling of detector to buffer even in the 3-MHz region I can hear "sideband cutting" set in on SSB signals. (See my "Super 3-in-9" receiver beginning with https://forums.qrz.com/index.php?th...pin-miniature-tube.897249/page-5#post-6801035 and rummaging around in that thread. With an IF tuning from 3 to 3.15 MHz, "clock rocks" (and for 30 m, 7-MHz ham-band crystals) can serve as a facile source of conversion crystals.)

A high-C regen is hard to pull before your ears overload. Main advantage to using a regen -- aside from the "cleanup" selectivity afforded by the oscillatory loop -- is that it requires much less AF gain than an equiv alent so-called "direct conversion" detector, variations on which include classical BFO-plus-diode and BFO-plus-product-detector heterodyne detection.

Omitting discussion of such dark-horse techniques as use of a gated beam tube for detection. That approach is high-gain -- with microphonics -- but no oscillatory-loop filtering.

 

Think not -- mult probs involved with tubes, which afford much lower gain than BJTs. Better to use a limiter.

 

@W9BRD: you are exactly right about selectively. I have satisfied myself with several experiments, going back years, that a good, properly designed, regen detector makes a better receiver for typical amateur use than most direct conversion type receivers, especially when wanting to use just valves/tubes.

I also agree that a soft limiter is much better than trying to implement AGC, for CW.

I also use toroid inductors for all my receivers. In most cases, higher Q is possible in real life than with air-cored because of the much reduced influence of metal, damping and unmanaged stray capacitance, etc.

As an experiment, again with a properly designed regen detector, and even operating as high as 2.5MHz, good enough single-signal (old school context) reception can be achieved by operating the detector below oscillation threshold and injecting an offset BFO signal. Obviously, that effect cannot be achieved when the detector is oscillating because the oscillation is centrally placed in the passband. It needs to be offset to fully appreciate the selectively available.

A 20m receiver of the Hikers type will definitely need a higher IF than you have currently. Even at 40m, in strong signal circumstances, you might have image breakthrough if only using a single tuned circuit on the input.

My 2.753 to 3.053MHz version is just good enough at 30m but I think even higher would be needed for 20m.

Philip

 

Lots of interesting stuff to digest, above.

Calling it a day, after operating several times this afternoon and evening. Had one very welcome QSO with Jim, @GM3ZMA, reassuring me that the signal really was getting out, although there was a lot of QRN and QRM.

RBN shows that the Peashooter is being skimmed all over the place, but at those very low dB levels it's not surprising that the QSO with Jim was the only one.

A reminder that we are dealing with a power output of about ... 700 mW.



73 de Martin, G3EDM

 

Luckily -- depending on how we look at it -- the HF bands are less-used now than decades ago, decreasing the likelhood of image interference generally, "depending."

"High side" conversion crystals and a tunable IF that tunes down while the frequency being received goes up can help with this in that higher-frequency bands -- where one's images would lie with high-side conversion crystals -- tend to be "daytime" bands while lower-frequency bands tend to be "nightttime." (Obviously there are exceptions; and of course at close range a transmitter being used for communication to anywhere can make for a troublesome image.)

My two "160 m as tunable IF" receivers both place the lower limit of a given cover ham band at 2.0 MHz and then tune down at 160 to tune up on the covered band. This works well with a simple single-tuned-circuit mixer front end (although I add that my parallel-resonance-tuned antenna tuner is always assisting somewhat), and works toward getting better behavior out of a regenerative detector because regens are generally better at lower frequencies. A 5.5-MHz conversion crystal brings down 80; 9 MHz, 40; 12 MHz, 30, and so on.

So now consider using a 3-MHz-region tunable IF -- say 3.0 to 2.9 MHz. A 6.5-MHz clock rock brings down 3.5 to 3.6 MHz; a 10-MHz clock rock brings down 7 to 7.1 MHz; a 13-MHz clock rock, 10 to 10.1 (more on this in a sec). Seemingly no clock rocks at 17 MHz, but second-harmonic mixing from an 8.5-MHz clock rock could suffice if signals at crystal f + 3 MHz (11.5-MHz region) do not obtrude. (Maybe there'll be some crossover responses; fun to calculate. I generally just do a quick build and try all the permutations to find clinkers.)

IMO one must be careful not to try to build an "all singing, all dancing" (to use Pat Hawker's variety-show metaphor for commercial receiver products built to appeal to so many that they're mediocre for specialist use) arrangement; that is, only build in capability, esp band coverage, you are likely to use. Also, as before, "the perfect is the enemy of the good": So what if a single-tuned-circuit mixer front end doesn't suppress a few strong spurii to below audibility -- are those signals really going to hamper comms? (Hint: On bands where the antenna-system floor is high anyway, build that single tuned circuit with "too much C"; that lowers tuned-circuit Z, sharpening tuning and generally reducing the level of signals applied to the mixer. Hint 2: This is why I have no problem using exactly 1.75 MHz as the fixed IF for my classical 80/40 band-imaging receivers: I never operate within earshot of the IF/BFO harmonics at 1.75, 3.5, 7 . . . , and anyway that quite distinctive-sounding signal acts as a band-edge/calibration marker.)

Because possible higher-frequency-band operation is under consideration for a/the next Peashooter, I therefore suggest a 3.0-to-2.9-MHz (or whatever more or less restrictive) IF span, using a fairly high value of C in the detector tuned circuit. (Assuming the tuning C across the entire detetector coil, you can experiment with tuning spans using the toroidal-core L/C/X calculators at kitsandparts.com -- even if you won't be using a toroidal coil in your design. Just use the calculator for, say, the T106-2 and get busy with cut and try.) The mixer plate feed can be via an RF choke, say 5 or 10 mH; then we lightly couple the detector tuned circuit to the mixer plate with -- experiment needed -- may 5 pF.)

Oh, and as for my 30-meter example, in which a 13-MHz rock brings down 10 to 10.1 MHz but misses 10.1 to 10.15: That's why you also build in a switch that gives coverage of 3.0 to 2.9 but also, say, 2.95 to 2.85 MHz or 2.9 to 2.8 MHz. A SPDT toggle could give three slightly different tunable-IF spreads: 3.1 to 3.0, 3.0 to 2.9, and 2.9 to 2.8; "center off" tunes the highest range and the two ON positions give the two lower ranges. These variations give you leeway in locating/trying suitable conversion crystals for ham-band and "bonus" coverage, such as for SWBC/standard frequency & time/weather.

BTW, why not build one's tunable IF for, say, one's lowest ham band of interest and then covert high bands down to that? Multi-aspect answer: (1) I like to build in transmit-signal monitoring, and such will be impossible on the tunable-IF band because we just can't keep a strong local signal at the detector fundamental out of the detector; (2) suitable conversion crystals that mesh with hamdom's classical harmonically-related bands have gotten harder to find; and (3) I think it design-better to keep a receiver's conversion topology constant from band to band if possible. (So, yes, I think the Drake 2-series receiver should not have used 80 as a tunable IF. Essentially all they did in using 80 m as the tunable IF is save one conversion crystal while complicating bandswitching and potential making the receiver more prone to overloading with the converter gain in play.)

 

Convert. @SM0AOM's unobtrusive Sunglasses And Fedora Guy will handle covert ops.

 

I am not sure, but having seen photos of Karl-Arne elsewhere, I think the Fedora Guy is Karl-Arne.

I like those photos because they give you a better idea just how large and cumbersome the Paraset is. Which reminds me, I need to get a proper photo of me operating the Peashooter. When you have a human in the photo, the scale becomes much clearer. It really is very small, for a tube rig. Lightweight ... not really, at 2kg!

73 de Martin, G3EDM

 

As a newcomer to operating /P, I have some questions about logging.

How does it work with LOTW? I assume there is a mechanism for telling it that you were operating /P, and I wonder whether it asks you where you were operating from?? I assume also that LOTW automatically excludes /P entries from any awards?

How would you recommend that I handle the paper logs? Right now, I have started a totally separate logbook for /P, partly because the smaller form factor (A5 size) is easier to deal with on the road. However, is that wise? Would it make more sense to transcribe it all into my main log book?

I apologize for simply asking these questions without even bothering to figure out the answers for myself (which can be one of the annoying things about forums like this one). It's just that I'm trying to deal with all sorts of non-radio stuff. Also, the "official" answers to questions about logging may differ from what people do in practice, so I'd like to know how people handle this in the real world.

73 de Martin, G3EDM

 

I operated the Peashooter "greyline" this morning for about one hour.



As you can see, my signal was spotted all over the place: UK, France, Germany, Spain, Sweden, Estonia, Slovenia, Hungary, Italy....

Maximum distance: more than 1,700km.

No contact, but that's not unexpected with 700mW, and the S/N numbers on RBN were weak.

(Also, the propagation forecast for 40m HF is "Poor").

I am still waiting for that elusive non-UK contact with the Peashooter.

73 de Martin, G3EDM

 

There is one odd thing, in my 40m version of the Hikers receiver. Yesterday I tuned a fairly strong signal calling CQ. A bit later I went upband about 20kHz and heard exactly the same signal, also calling CQ (but much weaker, basically in the background). Went back down 20kHz and there is was in its original strong form.

What is going on there? Obviously some kind of image, but I can't quite figure out the maths of it, how did it get offset by that relatively small amount?

(I remember getting a similar issue with my regenerative receiver years ago, and it was caused by some parasitic oscillation. I quenched that by adding grounded shields to a couple of the tubes.)

I should stress that in the Peashooter, this kind of odd "image" has been rare and only with strong signals. I have come across it maybe three times in the past couple of weeks.

Edited to add: I've also had very occasional BCB shortwave (or maybe even MW?) breakthrough, but it tends to be very short-lasting and pretty much lost in the background. I don't think it's happened more than a couple of times.

73 de Martin, G3EDM

 

Quenched -- good choice of words, because your detector may be superregenerating at 20ish kHz. A test for this would be to use another receiver to tune around its carrier to see if you can hear the spurs. Alternatively, tune around with the receiver receiving a strong local signal -- a crystal oscillator; throw one together if need be -- and see if you can hear similar "20-kHz away" spurs on the test signal. (A family of such sidebands should be present above and below the detector signal; in effect, the detector carrier is modulated by the suprerregenerative oscilation.)

Triode detectors are especially prone to superregeneration, but the behavior can also occur in screen-grid tubes. The triode detector in my Super 3-in-9 receiver goes into superregeneration if I increase regeneration far enough; in early experiments with the same triode used in the Hartley configuration instead of the Armstrong configuration now in use, the onset of superregeneration occurred not far above my standard regeneration level, so nix. When superregeneration occurs, the carrier of the oscillator "grows sidebands" spaced from the carrier (and from each other) by/at the superregen frequency, and decreasing in level with increasing distance from the carrier.

If you don't hear spurious sidebands on your detector's carrier, spurs on the received signal are the next likely cause. Another possibility is a spur in the 1R5 stage.

 

No, the clandestine operator in the pictures was Johnny/SM7UCZ(SK).
I do not even own a Fedora, but only a Tweed woolen cap.

Below I am "caught in the act" before my moustache went grey...

 

Finally getting round to recharging the batteries, first time since Saturday.

I've probably operated five or six hours since then, and no sign of the batteries flagging.

I can't remember for sure, but I think my calculations originally indicated a life of about 10 hours for the B+ batteries and they are the limiting factor (600 mA/H wired in series). The filament batteries add up to 11200 mA/H (four in parallel) which is substantial, in fact I could probably have left one or two of those four batteries at home, and saved a smidgeon of weight.

Basically this means that my hiking routine of recharging the batteries every night was excessive. OTOH there was always the possibility of a major propagation breakthrough and a very long session in the middle of the forest somewhere ... I wanted to be ready.

73 de Martin, G3EDM

 

I've just checked, and it looks like I can fit another five PP3 batteries horizontally along the top of the existing battery pack. The AA batteries for the filaments would be moved, probably to an existing empty cavity at the front of the briefcase, in front of the receiver. The alternative would be to move the AAs so they are mounted underneath the hinging lid, at the front. I don't much like that idea, I prefer the existing, clean design.

This would yield a B+ of about 160VDC and bring power output to nearer one watt. From a construction point of view, it has the great advantage of not requiring any alteration to the radio itself, but only to its (detachable) PSU. I may include a switch so that the older configuration of 15 PP3s remains available, leaving the horizontal five-battery rack out of service to save weight when required. However that is a potential failure point and, thinking about it, the extra weight of the five batteries and their connectors (about 150g) is not that big, in the grand scheme of things. It adds about 8% to the weight of the radio.

The challenge with the redesigned battery pack is that there is not in fact enough vertical room under the lid to place the horizontally mounted PP3s if (a) I use the same, fairly thick wooden batten as a base and (b) I use conventional battery holders, which have a base underneath the battery and therefore add to the thickness. I'll need to find a thinner, stiff material to replace the wood (carbon fibre??) and I think I'll have to design my own battery holders in such a way that they do not add to the thickness overall.

(I can hear you thinking: Why not use a different, slightly larger case? Well, it took me ages to find the maximum size of case that would be tolerable for long-distance hiking. Also, many hours were spent fitting the existing radio to the existing case ... I don't want to go through that again.)

Of course an increase in power from 700mW to perhaps 1000mW doesn't amount to a hill of beans in terms of dB gain, but every little bit probably helps.

I have various ideas in mind for a completely redesigned Peashooter with a much higher B+, using two rows of vertically mounted PP3 batteries but in a case no larger than the current one. Space would be saved by using subminiature tubes except for the transmitter's PA. This would pose major challenges in terms of "interface design" because, in fact, it's the human interface (capacitors, potentiometers, switches, meters, a Morse key) that turns out to present the biggest challenge in all these projects.

Whether such a radio, probably using a directly heated 2E30 PA tube, would work properly is an open question and @SM0AOM has already cast doubt about issues such as the internal resistance of the batteries when surge currents are required. So it would need to be tried out experimentally ("breadboarded") first. (That is a good procedure for all homebrew projects in any case.)

73 de Martin, G3EDM