My 120-mile Hike Carrying Homebrew Tube Transceiver

I first saw this thread because it was linked from a post in the GQRP Club's mailing list. I might have found it in my general browsing of QRZ, but probably not, even though I visit the QRZ forum fairly frequently. I am pretty sure there'll be others who'd be interested who never come to the QRZ forum even if they use QRZ for callsign lookups regularly.

 

750 mW from a 12A7 Colpitts crystal oscillator (that keys perfectly, BTW), subtitle "Why 80 for Regional QRP?""



Antenna is a low/bent/drooping 100-ft doublet center-fed with window line. Comfortable two-digit SNRs.

 

80m would be great at 2 in the morning in the fall when QRN is going down.
But big antennas on the trail during the day may not work the best for actual contacts.

 

The original Hiker's Portable by W6YBT (QST, September 1950) was monoband 80 metres. He included all sorts of elaborate arrangements to cover the entire band, in segments, with multiple crystals involved.

My version is for 40m and I do think 80m is a difficult band to operate in a lightweight, hiking context, given the length of antenna. In practice you'd often just settle for a long wire of some kind. W6YBT also used several wire counterpoises, and frankly I cannot imagine going to that trouble, in a "hiking" rig.

73 de Martin, G3EDM

 

The key to QRP success is frequency and band agility.
This cannot be repeated too often.

In the hands of a skilled operator, and perhapseven more important, skilled operators "at the other end", also low power "works".

However, today there is a quite marked threshold efffect when it comes
to low power operations, as the skilled operators who can use weak signals are becoming more and more scarce. It appeas that RBN does at better job than perhaps 80 % of the already few Morse operators.

"Back in the day", when radio amateurs as a whole were better qualified, and the noise levels were lower, it was not that uncommon to operate portable with only a watt or so from a battery one-tube oscillator transmitter and have regular QSO:s in the 300 to 500 km distance ranges on either 80 or 40 m.

Antennas mostly were random wires with a few counterpoises,
using pi-network matching. This form of antenna matching makes you not having to worry about SWR and related horrors.

Today, I am of the firm opinion that a minimm RF power output of 5 or 10 W to a reasonably good antenna is about minimum for successful operations when crystal-controlled, and 2-4 W when VFO controlled.

VFO control makes you able to move away from interference, and
to answer other stations calling CQ. Being able to work more bands makes it possible to follow the variation of the optimum frequency.

 

FYI never suggested DC RX. The hybrid is 1r5 converter set up as tuneable with simple ladder filter to product detector
(another 1r5) followed by 1u4 audio. Needs less audio gain, less prone to microphonics and is a true superhet single signal.

Two other benefits... higher IF with more selectivity and corresponding smaller coils needed.

An aside with the variable oscillator at the input converter end it makes the potential for common VFO for
single tuning for transceiver realizable. Thats a separate subject.

The solid state version of that is NE602-filter-NE602 is the common houseflies version of that. So the topology is
not exotic or complex.The proto seems to work well.

As to gated beam tubes or sheet beam tubes... nope! One they all consume great amounts of heater power, they are
prone to magnetic fields, most offer limited gain. Unsuitable in general especially for portable. I've worked with
7360 and 6AR8, ok for SSB balanced mod but not optimal from a power perspective.

An aside the idea is a CW RX no SSB and no AM. To do CW using regen requires it to be oscillating so the usual
tickler feedback Armstrong regen is now an Armstrong oscillator and to receiver CW its DC current is sampled for audio
making it a Autodyne converter where the oscillator is offset from the signal say 800hx for CW copy.

Since it has to oscillate to get CW tone why skip all the regen excess and go straight to a 1R5 as mixer and as a
product detector its oscillator section is tuned as needed and the grid input is the source signal. the plate circuit
is optimized for Audio output with low pass or peaked at 800hz (for CW).

As to AGC once you omit regen its easy and for CW its not commercial style AGC that requires a lot of gain for
good servo action its of the threshold exceeded pull back the gain aka ear saver. With smaller magnitude of
reduction available and needed the circuit requirements are small complexity minimal.

Last item... The large open coils buy nothing. A .5" slug tuned formfrom before WWII is far better or even
gasp toroids which are not that much younger. Key to high Q is less Resistance and that means less wire.
So powdered iron cores (solenoid or toroid) can easily exhibit better loaded Q than Airdux with less volume
and incidental coupling. This s important to portable as size and weight matters or only allowing more
room for battery.

 

Can you just confirm that this is a reference to the tank coil?

Thanks.

73 de Martin, G3EDM

 

some comments..

Weak signal ops are getting scarce, most I know are VHF and UP and their ear for CW are finely tuned
As are the radios for the job. I've found most commercial rigs are less effective due to internal noise.
Exceptions I have are my old (mid 1970s) Tentec Argonaut 505 (qrp) and Triton its 100W brother
and even at full volume the RX is is hard to hear...then you connect the antenna and Wow! The only
other not designed by me for HF is the Tentc Eagle, quiet RX hears exceedingly well. All three do
QSK unlike my TS440S/AT and IC7300.

>>Antennas mostly were random wires with a few counterpoises,
using pi-network matching. This form of antenna matching makes you not having to worry about SWR and related horrors.

While true for random antennas, based on a lot of historical research Pi networks were late in the game and introduced
to meet PTT/FCC rules. most designs were maybe 25 to 150 ohm range of match due to component size constraints.
Far more common was tapped tank or a tank with a variable (both mechanical and electrical) loading and both could
accept a wide range of "loads".

A note my favorite AMU is reversible L network matches everything from low impedance highly reactive loads to
EFHW wires directly.

However most of those also had minimally a plate meter for dip and load to ascertain correct tuning to get power
going to the radiator.

SWR is only relevant to a reference like 50 ohms. Modern gear mostly. I still use a a field strength meter
to tune the mobile whips.

As to VFO, flexibility and agility is always valuable. Its why my first grab portable is FT817, DC to daylight
capability and I'm VHF/UHF op. The kit I carry in the truck has antennas for all occasions and types
while compact.

The problem with that agility is complexity. Doing two bands isn't too bad but a larger spread takes much
more hardware and big gang switches. from actiol use 40/20 and maybe 15m does it.


Allison

 

Tank coil yes. Mine for 40M is about 30 turns of #26 and 10 turns secondary with a variable cap to ground
on the secondary. Matches less than 30 ohms to about 200 easily. A powdered iron slug will tolerate maybe
8-10W. It could be a toroid I tried a T50-2 and it worked well with 37 turns and 10 for the secondary.
I posted a picture of it on the toroid build thread for the trip.

Note since the overall length of wire is less the over all Q is higher. Measured Q unloaded is about 200.
The balance point is any ferrite lowers power handling to a point, but we in the less than 5W region so
not a factor.

Tapped coil is a pain and where the taps should be more so. Pi net wants bigger value caps and similar
size coil for wide range so no advantage there.

Allison

 

A product detector at the end of an IF strip is a d-c receiver.

BTW, have built very sharp ladder filters here, but strongly dislike them for general bandscanning and tuning around after CQs. And so then the non-trivial issue is how to switch between ladder filter and wide/none without compromising the filter's stopband rejection. Nah.

Because I find single-point/single-signal tuning unnecessary for general work, I tend to use an at-IF regenerative detector (because of its built-in loop filtering and relatively high AF output v RF input) and switchable passive AF low-pass filtering (daily driver AF amp/filter box has 500 Hz, 2.5 kHz, "through") for tougher going.

For general single-point tuning I'd settle for an SSB-width filter, but they are nontrivial to build for the low IFs I like to use. And I'm not going to use a commercial filter, even though I have a bunch at 3.395 and 3.18 MHz. Closest I get is the Snelgrove 6-kHz 1750-kHz I use to protect the NE602 product detector in one of my band-imagers; I do have cascaded 9-MHz KVG SSB filters in a solid-state modular covers-89-and-20 constellation that I keep just managing to not use for two-way work because it's, well, uninteresting despite the fact that it works well and likely has a 3rd order IMD DR around 90 dB.

And yes, as @G4HOJ has mentioned, using an oscillating regenerative detector immediately downstream of a crystal filter can be . . . comical. Time-delayed re-emission by the filter crystal(s) of the regen's carrier makes the detector perturb, and lock to, itself with unpleasant thumps and bangs, disallowing smooth/stable adjustment of the regen relative to filter center. Even as an operative fail such effects are fun to understand.

 

As to product detector vs DC RX. Your being pedantic as then the DSP in the IC7300 is DC RX as well.
OR the Argonaut 505 would be be as well at the end of the IF.

The desire for wide and narrow filter is long standing. Many CW ops ended up using audio filters
because they bought SSB bandwidth radios and often complained about it. Seems they didn't
get what the wanted only, what they purchased.

However two filters are doable with minimal effort like a pair of slide switches. Works in the HW101
and a lot of other radios.

Wide filters I do them, Not hard even at 3mhz once understood. Just good measurements, calculations,
and some care. And luck finding the right crystals if its a lattice filter. Ladder filters make it easier
but rocks matched for frequency and also for Q are a must.

Best low frequency not mechanical filter was built with 160xKhz (1600 adn 1602 lattice) crystals
I happened to get. it was 2.4Khz with a fairly even phase delay. Skirts at -60DB was 1.4:1.

However high if vs low if... I go high as lower phase delays are easy and convert up or down
its all the same to me. After working at GHZ closer to DC is not a big deal.

>>>using an oscillating regenerative detector immediately downstream of a crystal filter can be . . . comical.

I find that amusing as you need the regen to oscillate to RX CW. A lack of shielding and radiation
will make that a rat race situation.

Your choices and your bench. I"m not telling you to do it, only how I've done it without any of the
complaints listed.

Allison