A high crystal current: my own belief is that a crystal current of the order of 10s of mA harks back to the 1930s when the piezo-electric effect was effectively still in its infancy. With that, crystals would have been chunky affairs, not only capable of handling such currents, but the low crystal 'activity' (due to the type of cut, impurities, surface irregularities from the grinding process, etc.) would have required a circuit arrangement that would have
required such high currents to be produced in the crystal in the first place. A triode would have been the natural choice - with its attendant high Ca-g, providing a much higher feedback current than with a tetrode or pentode - and more power would have been required from that oscillator to drive successive stages (typically triodes), compared to a much more efficient all-pentode arrangement. Net result: a high crystal current by requirement; a crystal capable of handling such a high current.
However, the technique of inserting a light bulb (typically 2-v., 60 mA) in series with the crystal was still being advocated up until 1974 in the ARRL Handbook of that date and the RSGB H/B (4th. ed.) of 1971 (and other RSGB publications of about that period). However, I note that the latter does specifically refer to X-cut crystals using that bulb technique. But from the 1930s to the 1970 is a very long time in terms of radio / electronics technology, and during those later years, the X-cut was obsolete (superseded by more efficient cuts, such as AT and BT, amongst others). And that, in turn, generates one question: why the persistence of still referring to X-cut crystals in Amateur Radio publications during that later period? I can only suggest that Radio Amateurs are notorious for being extremely reluctant to dispose of
anything that one day just
might be useful (here, I speak from personal experience
): hence, one metric of any Radio Amateur's enthusiasm for his hobby was (and perhaps still is) the size of his 'junk box'!
An auxiliary comment is appropriate: I do not know the cut of the crystal used in my experiment. If I had had a
known X-cut type, I would have used that. And - just as an aside - I did try a 400 kHz HC6/U type: circuit operation was much the same; the crystal current was of insufficient amplitude for a reliable measurement to be made. (And that itself is highly relevant here.)
Finally (if it isn't too obvious), in this monologue, I have made a few assumptions - and a bit of 'educated guesswork'. So if anyone feels that they need to add or correct anything thus written, please do so: your comments will be welcome.
For me, the furtherance of Science is more important than the protection of one's ego.
Al. / Sept. 2, 2014 //
I have only had one reply to the question I posted on another forum:
Lawrence.
"There will be no current through the crystal as a result of cathode-grid diode action, as a crystal is an open circuit for dc. Cathode-grid diode action will, however, develop a dc bias level at the grid (charging the cathode-grid capacitance and discharging it through the resistor from grid to ground--the grid leak resistor).
But your main interest is the 6F6 as oscillator and RF current through the crystal.
The circuit you describe is a a tuned-plate, crystal-grid oscillator, an analog of the tuned-plate, tuned-grid oscillator in which the crystal acts (in its parallel-resonant mode) as a parallel-tuned circuit between the tube grid and ground. Positive feedback for TPTG/TPXG oscillator occurs through the 6F6's plate-to-grid capacitance--0.7 pF in a metal 6F6 (and its transmitting-rated alter ego, the 1621); more for a glass 6F6. The oscillator plate circuit must be tuned to a frequency close to, but above, crystal resonance for the phase of the feedback to be correct for oscillation.
RF voltage at the 6F6 grid makes the crystal vibrate. This vibration--and hence the stress on the crystal--is maximum at crystal resonance and increases with the RF voltage across the grid. Because the crystal is acting as a tuned circuit, and a tuned circuit excited by RF has RF current circulating through it, RF current flows through the crystal.
We should note that although a maximum for crystal current is generally specified, and that exceeding this maximum can result in crystal fracture, at least one source (James J. Lamb in his April 1937 QST article "A Practical Survey of Pentode and Beam Tube Crystal Oscillators for Fundamental and Second Harmonic Output") declared that RF current through a crystal is actually a derivative value and potentially misleading, as--as we know from an understanding of piezoelectricity--it's applied voltage that causes a crystal to deform (and possibly shatter) and not applied current. Lamb considered crystal current potentially misleading because away from resonance a crystal acts, in conjunction with its electrodes, as a quartz-dielectric capacitor. A frequency-nonselective current indicator, of which a panel lamp is one, in series with the crystal sums the effective values of all current through it and therefore may over-indicate if harmonic, LF or VHF parasitic-oscillation energy, or other off-crystal-resonance currents are present.
Lamb also mentions, however, that a lamp-in-series-with-the-crystal current indicator is therefore conservative as a means of indicating possible danger to the crystal.
In a TPXG oscillator, the greater the plate-to-grid capacitance of the oscillator tube, the greater the feedback and the generally greater the voltage across the crystal, and therefore the greater the current through it. Plate tuning therefore also affected the grid voltage, and hence the crystal stress. Potentially most dangerous to the crystal was its use in a TPXG oscillator used stand-alone as a transmitter--that is, coupled directly to an antenna. Loss of appreciable plate loading during tuning or as a result of sudden disconnection of the antenna could suddenly greatly increase feedback and destroy the crystal.
Not long after the series-lamp-as-crystal-current-indictator became popular, its users realized that the large resistance increase from cold to warm to hot in the lamp could compromise keying quality, resulting in yoop, a relatively slow frequency shift across multiple dots and dashes of Morse code elements. This led to builders equipping such lamps with a shorting switch; adjust the oscillator for crystal safety with the lamp unshorted, and then operate the transmitter with the lamp shorted for best keying.
Crystal current indication with a lamp would all but disappear rapidly after World War 2, as increasing attention to the crystal oscillator as a frequency standard rather than as a producer of appreciable RF driving power (and increasing use of RF power tubes that needed less drive for a given output power than prewar tubes) resulted in crystal oscillators generally being so low-power that the danger of crystal stress from overcurrent was all but banished by improved circuit designs.
And so this final note on the circuit you've found: Tiny crystals of modern manufacture (think HC-49), and even older HC-6 crystals, are likely to be destroyed when used in circuits like your 6F6-6L6 transmitter, especially crystals resonant at 7 MHz and above (because they're so thin). Even pre-World-War-2 we see, over and over, beginner's transmitter designs that went to great lengths to encourage the use of 160 or 80 meter crystals (relatively hard to fracture) rather than 40-meter ones."
Well, to non expert me, that sounded like a very expert reply and it was well written too. Clear and easy to understand so my thanks to the author.
Gary
Yes, 99.9% of what has been written there, I would agree with . . . for what
that's worth!
And I did draw some comfort from it . . . . .
Quote: at least one source (James J. Lamb in his April
1937 QST article) . . .
and
quote: Crystal current indication with a lamp would all but
disappear rapidly after World War 2 . . .
both of which have confirmed my earlier remarks, to wit: this idea of using a lamp for the purposes as described was decidedly obsolete by the 1960s, despite its continued appearance in Amateur Radio publications at that date - and after.
Al.
The xtal size on some of the American comms receivers seems to have shrunk around the time of WW2, the xtal in my HRO senior is a fair lump but in a later model (5t) the size is somewhat reduced ditto in the Hallicrafters SX 28 compared to the earlier SX 17.
Lawrence.
Yes Al, but what was the Lamp for?
Nowhere have I seen an explanation, or am I missing something?
Alan
Alan - I quote from the ARRL Handbook, 53rd. edition, 1976, page 137:
The power available (from a crystal oscillator) is limited by the heat that the crystal will stand without overheating. The amount of heating is dependent upon the r.f. crystal current, which, in turn, is a function of the amount of feedback necessary to provide proper excitation. Crystal heating - short of the danger point - results in frequency drift to an extent dependent upon
the way the crystal is cut.
► Note that last bit, highlighted in
blue (by me): I read that to be an indirect reference to an X-cut.
And on page 138:
Crystal current may be estimated by observing relative brilliance of a 60-mA dial lamp connected in series with the crystal. Current should be held to the minimum for satisfactory output by careful adjustment of excitation. Typical crystal currents should be in the vicinity of 40 mA.
On both those pages, there are various cct. diags. of crystal oscillators - but in none of them does the 'dial lamp in series with the crystal' actually appear.
OTOH, not all of the ccts. thus shown are valved-based; some use bipolar transistors - and at least one is shown using a F.E.T.
Personally, I reckon several chunks of those extracts were simply copied from the previous edition of that book, which were copied from an earlier edition, which were copied . . . .
ad nauseum . . . and possibly right back to edition #1 !
In which case, the editorial attitude must have been "If in doubt, leave it in".
Al. / Sept. 3, 2014 //
A subsidiary thought.
Going back in time to when this whole issue of "don't overheat your crystal!
" was so obviously such an important consideration in the design of crystal oscillators, you'd have thought that the crystal manufacturers would have stated a maximum amount of r.f. power / r.f. current that a given crystal could withstand - similar to the wattage rating of a Zener diode, for example. But I've never come upon such a rating, either by a manufacturer explicitly, or implicitly in a circuit design.
Has anyone else?
Al. / Sept. 3, 2014 //
Just for reference: extract from the RSGB H/B, 4th. edition (1971).
Characteristics of quartz crystals.
Type of cut: X Y AT BC BT GT
Normal freq. range, MHz: 1-5 1-10 0.5-8 1-20 1-20 0.1-0.5
Temp. Co-eff. ( Hz/ MHz/ °C): -20 +75 0 -20 0 0
The temp. co. of the AT and BT cuts, given as zero, is only true at certain temperatures; at other temperatures, likely to be a small +ve or -ve value. Only the GT cut has a true zero temp. co.: 0 °C → 100 °C.
-----------------------
(Presumably, the 'normal freq. range' refers to fundamental freq. operation and not an overtone).
It should also be borne in mind that that data was (presumably) up-to-date in 1971.
Al.
You're becoming nocturnal young Al
Thanks for the explanation. So, is the Positive Temperature Coefficient of the Lamp of no consequence, ie it's just a visual Indicator?
Alan