[HBR] Tube history/usage summary?

[Walt Hutchens]

George said:
> I know transistors like the back of my hand, but the experience of:
> which tubes are newer/older, what tubes are better for this or that,
> what tube characteristics are really important or not, what tube is
> subject to parasitics more or less, et al ... is all a part of
> history that I¹ve missed out on over the years.

Tim's response is really excellent.

There were two main trends in vacuum tube design over the years:

1. Improved accuracy in construction, allowing for smaller structures,
closer element spacings, and so on.  These allowed higher gain and --
since the upper frequency limit of a tube is established by the size
of its element structure in wavelengths -- successful operation at
steadily higher frequencies.

(Wartime designs look like an exception: the 2C40, 2C39, etc.,
'lighthouse' tubes will perk along at around 3000 Mcs. But these tubes
succeeded by cheating: The elements are flat and are stacked up, held
in place by being fused to rings of glass. This allowed precise
parallelism and extremely close spacings -- but the structure was
horrifically expensive to make and sockets and circuits to use it,
equally costly. Only the military and aviation made extensive use of
these devices.)

2. Better understanding of tube design principles and operation
allowing the design of better tubes for specific applications. For
example, the 6EH7 -- one of the last TV IF pentodes -- is hands down a
better RF/IF tube for high gain WITH outstanding strong signal
performance, and particularly so in an AVC-controlled stage. The last
generation of power tubes for SSB linear amplifier service are far
better than the tubes first adapted to this job -- and so on.

There were many specific innovations that had substantial impact: The
higher emission cathode coatings (of about 1960 I think?) made
possible more tube with less heater power:  The 35GL6 is identical to
the 50C5 but with about 1/4 less power.

It's helpful to understand how the numbers were assigned to U.S.
receiving tubes.  The first digit, of course, was the filament
voltage.  Tubes numbered '6' have 6.3 volt filaments.

Within a filament voltage, the first letter goes in sequence: The 6A6
is an OLD tube the 6E5 came later, the 6L6 later still ... these are
all 1930's tubes, many of them with large (old) five to seven pin
bases.

HOWEVER ... the last digit -- think of this as the number of elements,
thus a '4' is a triode with a separate cathode -- qualifies the
letter. So tubes with a different number of elements might get
assigned a seemingly much earlier number: The 6C4 is a miniature 7-pin
base triode, an early post-war type.

When the number-letter-number assignments ran out, a second letter was
added right after the number.  The 6AK6, for example, is another early
post-war type.  Development of new types pretty well ran out by the
time this added letter hit mid-alphabet: The latest number that comes
to mind is the 6ME8 -- a 'beam deflection' type, developed for TV
color demodulator service but useful in mixing applications.

There are some 'wrinkles': Rectifiers got high letter assignments, the
letter 'S' as a first letter was used in the late 1930's, and probably
more that don't come to mind.

Later types are simply better tubes -- higher gain, lower noise,
better linearity, etc. But that doesn't necessarily translate to "you
can build better equipment with later types." You can almost always
build equivalent equipment with fewer tubes using the later types,
especially if you can use dual triodes or triode-pentodes in place of
separate tubes.  If you're building for VHF, yes, the later types are
unmistakably better, a conventional superhet (RF-mixer-IF-IF-detector
...) design will be better with later tubes, but who says you have to
be conventional?

Because later tubes have higher gain, technique issues that wouldn't
matter for earlier tubes must be gotten right. VHF parasitic
oscillations aren't a defect in the tube, they are a consequence of a
particular combination of a tube and the physical layout of the
circuit in which that tube is used. If you take a receiver designed
around 6SK7s (remote cutoff pentodes of late 30's vintage) and
substitute 6EH7s (same, 1960's vintage) you will probably either have
to take steps to improve the physical circuit, or use the better tubes
at much below their full gain.

WHAT steps? Well, that will depend on the design and physical
construction. The DC voltages may need to be changed somewhat. The
6EH7 should be compatible with the AVC voltage used with 6SK7s. Better
bypassing might be required, perhaps shorter leads (not always
possible), suppression of VHF parasitics ... If three 6SK7 stages are
replaced, there will be far too much overall gain, possibly leading to
a multi-stage instability.

I'm about 3/4 through building a variation on the W6TC receiver using
later tubes. The thing gave me absolute fits with RF-stage instability
until I returned the 'cold' ends of the coils directly to the
respective rotor contacts on the tuning cap.  Looking back, this
should have been obvious but it took me a month of occasional head
scratching to discover.

In beginning to build 'vintage-type' equipment, I think one is wise to
build at least a couple of established designs with as few changes as
possible. The ARRL Handbook designs are probably the best-tested of
the simple receivers and transmitters. (The W6TC 'HBR' series is the
major exception -- NO receiver was more widely built or described in
the ham literature. But this is a lot of receiver for a beginner to
tackle.) When a tested design is written up by an experienced author,
the construction points that must be gotten right are generally
explained in the text.

(Finding the parts to duplicate an established design, however, can be
a real challenge.  And when you start substituting, you lose some
fraction of the value of 'established' -- maybe most of it.  Changing
from one brand of 455 kcs IFTs to another contemporary brand probably
won't cause trouble but if you decide to substitute IFTs of a
different frequency, either with or without adjustments to the tuning,
all bets are off for those stages.)

Because the best tubes were designed for TV applications right at the
end of the vacuum tube era when ham designers were running open
mouthed after transistor designs, few well-tested ham designs exist
for these tubes. Much of what there is, is one-off: A smart guy built
one and wrote it up for CQ or QST.  This is both bad and good news:
The bad is that there's not much equipment you can simply copy, but
the good is that there's a very interesting challenge for the
determined home constructor.

The bulk of ham vacuum tube design followed well-worn trails, based on
limited understanding of the issues. Commercial ham sets mostly
reduced-cost variations on commercial comm equipment and nearly all of
them had strictly limited design budgets. (Collins was an exception,
National also, to a lesser extent.) Hams who 'rolled their own'
generally did so starting from the commercial design base. I think the
best ham equipment designs using vacuum tubes were never built.

Walt Hutchens
KJ4KV