On 4/22/2021 5:04 AM, piglet wrote:
On 22/04/2021 5:00 am, bitrex wrote:

There's this patent from '66 that describes a solid-state thyratron
replacement:

https://patentimages.storage.googleapis.com/42/6a/0d/6e35c415a3f0e0/US3293449.pdf


It uses a couple of SCRs. It's to replace high-power thyratrons but I
don't see why it couldn't be scaled-down.

Thanks Bitrex, that patent even mentions a tunnel diode and UJT which
are now very exotic devices. The struggle was to emulate the high
impedance thyratron grid. It seems to me (I have never seen a thyratron)
that a modern depletion mode mosfet would make a good front-end to a SCR
in that a negative grid bias is required to keep the device
non-conducting (if I understand thyratrons correctly).

Well Fig 2 in the patent doesn't look like it requires anything exotic,
the exotics look to be included in the "further embodiment of the invention"

Here is a conceptual sketch:

https://www.dropbox.com/s/al1b8hj07i7jfpq/SED_Thyratron_Idea.pdf?dl=0

Possible fets could be BSS126 or LND150 and SCR with sensitive gates
X0402NF or BT149G.

piglet

Yeah, depletion mode MOSFET seems a plausible modern take on it. Only
thing is a real gas thyratron the anode current is controlled by the
control grid current when it's in the low impedance ionized state. In
many switching applications though I expect this doesn't matter too much
it doesn't matter what the impedance is in that state exactly, just so
long as it's low enough to reliably fire the relay or whatever.

And there doesn't need to be an analog to the screen grid, when the tube
is used like a switch to ground it's usually connected to the cathode
and just prevents the thyratron from firing until the positive-going
pulse on the control grid is seen. You can see in a 2D21 a box-like
structure surrounding all the other electrodes and I believe that whole
structure is connected to the screen grid pin.