[Big Bad Bob]

On 10/07/16 16:22, Patrick Turner so wittily quipped:
I think the FIRST thing anyone should do to plot the Ea vs Ia average characteristics of a tube is to have a real good look at the schematics used in 1955, well before solid state and digital.
It is most remarkable that in 1955, curves were produced which gave us a fairly good guide to use for load line analysis.

Why do we not build on the skills of those who preceded us?

in my own personal opinion, I'd like to build something automated, using
some kind of computing device. my device of choice, a microcontroller.
It's really the same BASIC idea, apply a voltage and measure the
current, but with automation.

in the case of the device mentioned online, it's using patch cables. I
would rather use electronic switching to do the same thing, but no
relays or klunky mechanical switches. And switching regulators that can
be controlled via the CPU to create the various voltages.

I see that gadget at that link to the site for a curve tracer to be something absolutely could not be repaired and serviced if one of the 2 billion devices went phut due to high volts from tube tests.

you probably have a point, yeah. A properly designed system would be
hardened and able to be short-circuited without damage.

I see a toy, not a tool.

it wasn't my preference for a design, but it _does_ "work"

To test output tubes for class A you need to have Ea max up to 1,000V.

that's a very good point. And possibly higher. hardest part is getting
devices that behave well at those voltages. It's likely that a IGBT
device would do it, though, driven by opto-isolator and/or AC signal
with blocking capacitors. I've thought of both of these possibilities,
actually. Opto-isolator is probably the better solution. AC+blocking
capacitor might work well for the gate drive, though, to provide
sufficient voltage to turn the thing on [via an optoisolator again].

If you have an SE 6550 with Ea = 500V, Ia = 50mA, then the anode swing is +/- 450Vpk with a load of say 9k0.

right, and you want to curve-trace a tube through its entire operating
range if possible [without doing damage to the tube, naturally] to
verify that it's working properly.


My solution to not having a tube tester or a curve tracer or having reliable old data curves, or curves for newer Russian tubes is to set up a tube in a dummy class A SE amp and measure THD and Ra then calculate gm g1 and and gm g2 and µ g2 and µg2 to obtain the average properties at the idle condition - and its these properties we employ for listening to 90% of music.

interesting point, having a set of SE output transformers to test output
signals with. "dial-an-impedence" and then let 'er rip. Expensive,
though. I'm looking for low overall cost.

I have done a couple of new web pages to explain all this more better, but having some trouble with page uploads and getting rid of an unwanted tracking program that self-installed.

TRACKING program that self-installed? that sounds kinda, bad... [like a
virus infection]

[FYI I hand-code all of my HTML, don't use any of those web page 'admin'
systems, rarely use scripting if it's not doxygen-generated
documentation... have never seen anything 'injected' at all]