On Sun, Mar 24, 2024 at 7:56?PM Richard Knoppow <1oldlens1@...> wrote:
? ? It seems to me that the Tube Collector's group may have information
on how to distinguish the actual manufacturer, done by means of
construction, as you point out, and sometimes by markings. Some of the
markings may be subtle. I am not sure if this stuff is on their web site
or is available to non-members.
On 3/24/2024 1:56 PM, Clyde K2UE wrote:
> I had the good fortune to spend a good part of my working years as an RF
> Engineer at RFCOM and Scientific Radio Systems, and designed and
> supported commercial SSB gear with both "transmitting tubes" (6146,
> 8122, 3-400Z, 3-100Z, etc) and sweep tubes.? One of the things you have
> to do in a commercial manufacturing environment is evaluate and approve
> vendor changes, including tubes.? One of the first things you learn is
> that the name stenciled on the tube is not always the actual
> manufacturer, and there were rude surprises when major manufacturers
> decided not to run a give type (permanent or temporary) and instead
> bought unmarked tubes from a competitor and mark them as theirs.? We
> learned to discern details of internal and external construction as a
> clue to this sort of substitution.
>
> So since we can't really count on the brand name, how would a commercial
> PA designer approach this:
> 1.? First match tubes by internal construction details that give away
> the actual factory:? mica shapes, plate contours, grid leads and welding.
> 2.? Among the most similar I would want to first match by
> transconductance using a mutual conductance tester.? If the anode
> current change for a given grid change matches they will share the load
> on a large signal basis.
> 3.? Last I would match by idle current, so they start at similar
> currents and hence will track each other for best load sharing and
> linearity.
>
> My opinion, but based on actual experience with many thousands of tubes,
> YMMV.
>
> ______________________________________________________________
> Clyde Washburn, K2UE
> ClydeWashburn1@... <mailto:ClydeWashburn1@...>
