0
$\begingroup$

I'd like to build a Colpitts (or Hartley) oscillator based one-tube CW transmitter, and I'd like to use a VFO circuit rather than crystal control (in part because of the expense and limited availability of suitable crystals in the 80m and 40m CW bands). One significant limitation for these oscillator types, however, is that it's relatively difficult to control chirp and drift to meet FCC spectral purity requirements that didn't exist when these transmitters were common.

I've understood that much of both forms of frequency deviation is due to thermal effects, so it occurs to me that a relatively simple temperature control system (something along the lines of a soldering iron element controlled by a bimetal thermostat, all enclosed in a small box) might be applied to limit thermal excursion and stabilize frequency.

Is this practical, and if so, which components need to be temperature stabilized (coils, capacitors, or the oscillator tube itself)?

Improve this question
$\endgroup$
6
  • $\begingroup$ Well, drift is thermally related. Chirp is not. $\endgroup$
    – Mike Waters ♦
    Sep 17 at 22:09
  • 1
    $\begingroup$ @MikeWaters That looks like the tl;dr for an answer. BTW, I removed the "chirp" tag, as that refers to the software for programming computer controlled radios and isn't applicable to vacuum tube analog equipment. $\endgroup$
    – Zeiss Ikon
    Sep 18 at 11:03
  • $\begingroup$ hey, "expense and availability of 40m band crystals": we're talking 50 ct to 1.36€ here, available for shipping pretty much worldwide; I think it's OK to say if you just want to build a VFO circuit, that's fair and somewhat exciting :) You're not doing vacuum tube designs in 2023 because you're aiming for cheap or simple, and that's OK. $\endgroup$
    – Marcus Müller
    Sep 19 at 18:06
  • $\begingroup$ @MarcusMüller If you want the larger crystals that can run directly in a tube oscillator, however, they're limited in availability and more so in frequency selection. But yes, I want to build a VFO so I can tune the transmitter frequency more or less the same way I'll tune the (probably tube regen) receiver... $\endgroup$
    – Zeiss Ikon
    Sep 20 at 11:01
  • $\begingroup$ @ZeissIkon :) yeah as said, I think that's cool! $\endgroup$
    – Marcus Müller
    Sep 20 at 12:28

1 Answer 1

Reset to default
3
$\begingroup$ 
[...] I'd like to use a VFO circuit rather than crystal control

In a one-tube transmitter?...that's going to be very difficult.

Drift is certainly temperature-related. Thermal time constant is sometimes slow enough that re-tuning can keep you on the same frequency, especially for the larger components used in tube circuits.
Since the frequency-determining L-C components of the oscillator must be well-shielded, the mass of metal pretty much ensures slow drift. Temperature-compensating capacitors (or circuits) can be added to compensate for drift.
Huff-n-puff frequency-locked loops could be added to stabilize frequency, but only if the oscillator was on continuously.

Chirp on the other hand is mostly determined by biasing of the oscillator, along with tube characteristics. A triode oscillator would be nearly impossible to tame, while a pentode might be merely very difficult.
Again, the L-C components that affect frequency should reside in a dedicated shielded enclosure to prevent radiated power from feeding back. Nevertheless, in a one-tube circuit there exists a feedback path through the tube's inter-electrode capacitance that affects frequency, causing chirp as RF power builds to maximum.

A one-tube transmitter doesn't allow the commonly-used solution of a separate oscillator followed by a power stage. The oscillator runs continuously, while the power stage is keyed.
To mitigate chirp, the key might switch between a dummy load and antenna, with the tube oscillating 100%, but this would only be possible for very a low power transmitter and at low frequencies where the back-wave is well-attenuated.

Improve this answer
$\endgroup$
2
  • $\begingroup$ What sort of single-tube TX will get out of QRP power range? I expect to get between 1 and 10 W, depending what plate voltage I come up with. And yes, I'll most likely use a pentode; I have a couple of those on hand. $\endgroup$
    – Zeiss Ikon
    Sep 18 at 14:47
  • $\begingroup$ It occurs to me, if the LC tank were driven by one of the grids instead of the plate, it might be possible to let it run continuously and control output with a later grid. I'll ask another question for that... $\endgroup$
    – Zeiss Ikon
    Sep 18 at 16:13

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .